part-3-predictions.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Optimizing App Offers With Starbucks\n",
    "### Notebook 3: Predictions\n",
    "\n",
    "## Data Preprocessing"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "import numpy as np\n",
    "import os\n",
    "import io\n",
    "\n",
    "import matplotlib.pyplot as plt\n",
    "import matplotlib\n",
    "%matplotlib inline\n",
    "\n",
    "import seaborn as sns\n",
    "sns.set(rc={'figure.figsize':(12,8)})\n",
    "\n",
    "import sagemaker\n",
    "from sagemaker import get_execution_role\n",
    "from sagemaker.amazon.amazon_estimator import get_image_uri\n",
    "from sagemaker.predictor import csv_serializer\n",
    "\n",
    "import sklearn.model_selection\n",
    "\n",
    "import warnings\n",
    "warnings.filterwarnings(\"ignore\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Custom colors that we will use in graphs:\n",
    "custom_colors = ['#006241', '#84233C', '#1E3932', '#9D5116', '#E44C2C']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>possible_reward</th>\n",
       "      <th>spent_required</th>\n",
       "      <th>duration</th>\n",
       "      <th>offer_type</th>\n",
       "      <th>offer_id</th>\n",
       "      <th>short_offer_id</th>\n",
       "      <th>duration_hrs</th>\n",
       "      <th>channel_email</th>\n",
       "      <th>channel_mobile</th>\n",
       "      <th>channel_social</th>\n",
       "      <th>...</th>\n",
       "      <th>days_of_membership</th>\n",
       "      <th>event</th>\n",
       "      <th>hours_after_start</th>\n",
       "      <th>money_spent</th>\n",
       "      <th>money_gained</th>\n",
       "      <th>event_offer_completed</th>\n",
       "      <th>event_offer_received</th>\n",
       "      <th>event_offer_viewed</th>\n",
       "      <th>event_transaction</th>\n",
       "      <th>event_auto_completed</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>10.0</td>\n",
       "      <td>10.0</td>\n",
       "      <td>7.0</td>\n",
       "      <td>bogo</td>\n",
       "      <td>ae264e3637204a6fb9bb56bc8210ddfd</td>\n",
       "      <td>A</td>\n",
       "      <td>168.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>...</td>\n",
       "      <td>527</td>\n",
       "      <td>offer received</td>\n",
       "      <td>408</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>10.0</td>\n",
       "      <td>10.0</td>\n",
       "      <td>7.0</td>\n",
       "      <td>bogo</td>\n",
       "      <td>ae264e3637204a6fb9bb56bc8210ddfd</td>\n",
       "      <td>A</td>\n",
       "      <td>168.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>...</td>\n",
       "      <td>527</td>\n",
       "      <td>offer viewed</td>\n",
       "      <td>408</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>10.0</td>\n",
       "      <td>10.0</td>\n",
       "      <td>7.0</td>\n",
       "      <td>bogo</td>\n",
       "      <td>ae264e3637204a6fb9bb56bc8210ddfd</td>\n",
       "      <td>A</td>\n",
       "      <td>168.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>...</td>\n",
       "      <td>381</td>\n",
       "      <td>offer received</td>\n",
       "      <td>168</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>10.0</td>\n",
       "      <td>10.0</td>\n",
       "      <td>7.0</td>\n",
       "      <td>bogo</td>\n",
       "      <td>ae264e3637204a6fb9bb56bc8210ddfd</td>\n",
       "      <td>A</td>\n",
       "      <td>168.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>...</td>\n",
       "      <td>420</td>\n",
       "      <td>offer received</td>\n",
       "      <td>576</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>10.0</td>\n",
       "      <td>10.0</td>\n",
       "      <td>7.0</td>\n",
       "      <td>bogo</td>\n",
       "      <td>ae264e3637204a6fb9bb56bc8210ddfd</td>\n",
       "      <td>A</td>\n",
       "      <td>168.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>...</td>\n",
       "      <td>420</td>\n",
       "      <td>offer completed</td>\n",
       "      <td>684</td>\n",
       "      <td>54.61</td>\n",
       "      <td>10.0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 26 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "   possible_reward  spent_required  duration offer_type  \\\n",
       "0             10.0            10.0       7.0       bogo   \n",
       "1             10.0            10.0       7.0       bogo   \n",
       "2             10.0            10.0       7.0       bogo   \n",
       "3             10.0            10.0       7.0       bogo   \n",
       "4             10.0            10.0       7.0       bogo   \n",
       "\n",
       "                           offer_id short_offer_id  duration_hrs  \\\n",
       "0  ae264e3637204a6fb9bb56bc8210ddfd              A         168.0   \n",
       "1  ae264e3637204a6fb9bb56bc8210ddfd              A         168.0   \n",
       "2  ae264e3637204a6fb9bb56bc8210ddfd              A         168.0   \n",
       "3  ae264e3637204a6fb9bb56bc8210ddfd              A         168.0   \n",
       "4  ae264e3637204a6fb9bb56bc8210ddfd              A         168.0   \n",
       "\n",
       "   channel_email  channel_mobile  channel_social  ...  days_of_membership  \\\n",
       "0            1.0             1.0             1.0  ...                 527   \n",
       "1            1.0             1.0             1.0  ...                 527   \n",
       "2            1.0             1.0             1.0  ...                 381   \n",
       "3            1.0             1.0             1.0  ...                 420   \n",
       "4            1.0             1.0             1.0  ...                 420   \n",
       "\n",
       "             event  hours_after_start money_spent money_gained  \\\n",
       "0   offer received                408         NaN          NaN   \n",
       "1     offer viewed                408         NaN          NaN   \n",
       "2   offer received                168         NaN          NaN   \n",
       "3   offer received                576         NaN          NaN   \n",
       "4  offer completed                684       54.61         10.0   \n",
       "\n",
       "   event_offer_completed  event_offer_received event_offer_viewed  \\\n",
       "0                      0                     1                  0   \n",
       "1                      0                     0                  1   \n",
       "2                      0                     1                  0   \n",
       "3                      0                     1                  0   \n",
       "4                      1                     0                  1   \n",
       "\n",
       "   event_transaction  event_auto_completed  \n",
       "0                  0                     0  \n",
       "1                  0                     0  \n",
       "2                  0                     0  \n",
       "3                  0                     0  \n",
       "4                  1                     0  \n",
       "\n",
       "[5 rows x 26 columns]"
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "csv_file = 'data/clean_df.csv'\n",
    "clean_df = pd.read_csv(csv_file)\n",
    "\n",
    "clean_df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "['possible_reward',\n",
       " 'spent_required',\n",
       " 'duration',\n",
       " 'offer_type',\n",
       " 'offer_id',\n",
       " 'short_offer_id',\n",
       " 'duration_hrs',\n",
       " 'channel_email',\n",
       " 'channel_mobile',\n",
       " 'channel_social',\n",
       " 'channel_web',\n",
       " 'gender',\n",
       " 'age',\n",
       " 'person',\n",
       " 'became_member_on',\n",
       " 'income',\n",
       " 'days_of_membership',\n",
       " 'event',\n",
       " 'hours_after_start',\n",
       " 'money_spent',\n",
       " 'money_gained',\n",
       " 'event_offer_completed',\n",
       " 'event_offer_received',\n",
       " 'event_offer_viewed',\n",
       " 'event_transaction',\n",
       " 'event_auto_completed']"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "list(clean_df.columns) "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Drop the \"unknown\" customer data\n",
    "\n",
    "See Data Exploration notebook for more info about \"unknown\" customers."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "count    209030.000000\n",
       "mean         53.318323\n",
       "std          17.724360\n",
       "min          18.000000\n",
       "25%          40.000000\n",
       "50%          54.000000\n",
       "75%          66.000000\n",
       "max         101.000000\n",
       "Name: age, dtype: float64"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "clean_df = clean_df[clean_df.age != -1]\n",
    "clean_df.age.describe()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Drop the columns that will not be used for modeling"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>person</th>\n",
       "      <th>gender</th>\n",
       "      <th>age</th>\n",
       "      <th>income</th>\n",
       "      <th>offer_type</th>\n",
       "      <th>event_offer_received</th>\n",
       "      <th>event_offer_viewed</th>\n",
       "      <th>event_offer_completed</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>173100</th>\n",
       "      <td>03b326a99f5345df8b4d21ed26bdb0da</td>\n",
       "      <td>F</td>\n",
       "      <td>80</td>\n",
       "      <td>85000.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>195076</th>\n",
       "      <td>190fcd45fd074de8b41531fc6ffa6f42</td>\n",
       "      <td>M</td>\n",
       "      <td>26</td>\n",
       "      <td>33000.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6029</th>\n",
       "      <td>a89e75b62ad648d68d4160653f36d901</td>\n",
       "      <td>M</td>\n",
       "      <td>44</td>\n",
       "      <td>65000.0</td>\n",
       "      <td>bogo</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                  person gender  age   income offer_type  \\\n",
       "173100  03b326a99f5345df8b4d21ed26bdb0da      F   80  85000.0        NaN   \n",
       "195076  190fcd45fd074de8b41531fc6ffa6f42      M   26  33000.0        NaN   \n",
       "6029    a89e75b62ad648d68d4160653f36d901      M   44  65000.0       bogo   \n",
       "\n",
       "        event_offer_received  event_offer_viewed  event_offer_completed  \n",
       "173100                     0                   0                      0  \n",
       "195076                     0                   0                      0  \n",
       "6029                       1                   0                      0  "
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "columns_to_keep = ['person',\n",
    " 'gender',\n",
    " 'age',\n",
    " 'income',\n",
    " 'offer_type',\n",
    " 'event_offer_received',\n",
    " 'event_offer_viewed',\n",
    " 'event_offer_completed']\n",
    "\n",
    "clean_df = clean_df[columns_to_keep]\n",
    "clean_df.sample(3)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Prepare data labels - \"no view\", \"no order\", \"possible order\""
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>person</th>\n",
       "      <th>gender</th>\n",
       "      <th>age</th>\n",
       "      <th>income</th>\n",
       "      <th>offer_type</th>\n",
       "      <th>event_offer_received</th>\n",
       "      <th>event_offer_viewed</th>\n",
       "      <th>event_offer_completed</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>15088</th>\n",
       "      <td>6da8765ed47745959fd67865d3799206</td>\n",
       "      <td>F</td>\n",
       "      <td>82</td>\n",
       "      <td>84000.0</td>\n",
       "      <td>informational</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>15134</th>\n",
       "      <td>6e084f6ef50c46f0af01e02060a7c1a0</td>\n",
       "      <td>M</td>\n",
       "      <td>39</td>\n",
       "      <td>56000.0</td>\n",
       "      <td>discount</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>33271</th>\n",
       "      <td>f10d90acb03b43e094f755e607388657</td>\n",
       "      <td>F</td>\n",
       "      <td>63</td>\n",
       "      <td>113000.0</td>\n",
       "      <td>discount</td>\n",
       "      <td>3</td>\n",
       "      <td>2</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                 person gender  age    income     offer_type  \\\n",
       "15088  6da8765ed47745959fd67865d3799206      F   82   84000.0  informational   \n",
       "15134  6e084f6ef50c46f0af01e02060a7c1a0      M   39   56000.0       discount   \n",
       "33271  f10d90acb03b43e094f755e607388657      F   63  113000.0       discount   \n",
       "\n",
       "       event_offer_received  event_offer_viewed  event_offer_completed  \n",
       "15088                     1                   1                      1  \n",
       "15134                     2                   2                      1  \n",
       "33271                     3                   2                      1  "
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df = clean_df.groupby(['person', 'gender', 'age', 'income', 'offer_type'], as_index=False).sum()\n",
    "df.sample(3)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>person</th>\n",
       "      <th>gender</th>\n",
       "      <th>age</th>\n",
       "      <th>income</th>\n",
       "      <th>offer_type</th>\n",
       "      <th>event_offer_received</th>\n",
       "      <th>event_offer_viewed</th>\n",
       "      <th>event_offer_completed</th>\n",
       "      <th>label</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>15224</th>\n",
       "      <td>6ea759e2e1234242a48a575597924934</td>\n",
       "      <td>F</td>\n",
       "      <td>52</td>\n",
       "      <td>84000.0</td>\n",
       "      <td>discount</td>\n",
       "      <td>2</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>possible order</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>33807</th>\n",
       "      <td>f5112242df984403bfb4176d4487aef3</td>\n",
       "      <td>F</td>\n",
       "      <td>96</td>\n",
       "      <td>72000.0</td>\n",
       "      <td>discount</td>\n",
       "      <td>3</td>\n",
       "      <td>3</td>\n",
       "      <td>2</td>\n",
       "      <td>possible order</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>30111</th>\n",
       "      <td>da50429e4d43437eb8073e4264e686a5</td>\n",
       "      <td>M</td>\n",
       "      <td>61</td>\n",
       "      <td>82000.0</td>\n",
       "      <td>bogo</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>possible order</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1209</th>\n",
       "      <td>093f0c09cc7242228edbb0e52ca14af6</td>\n",
       "      <td>M</td>\n",
       "      <td>36</td>\n",
       "      <td>35000.0</td>\n",
       "      <td>discount</td>\n",
       "      <td>3</td>\n",
       "      <td>3</td>\n",
       "      <td>3</td>\n",
       "      <td>possible order</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>26342</th>\n",
       "      <td>bd777dba9c784ff8aa96c2a4531fba1f</td>\n",
       "      <td>M</td>\n",
       "      <td>84</td>\n",
       "      <td>93000.0</td>\n",
       "      <td>bogo</td>\n",
       "      <td>3</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>possible order</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25459</th>\n",
       "      <td>b702b01e1767462aa447431c9e8a07f9</td>\n",
       "      <td>M</td>\n",
       "      <td>57</td>\n",
       "      <td>110000.0</td>\n",
       "      <td>bogo</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>possible order</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5533</th>\n",
       "      <td>2a070c1a63e348fda6f9772df48f4c85</td>\n",
       "      <td>M</td>\n",
       "      <td>73</td>\n",
       "      <td>36000.0</td>\n",
       "      <td>discount</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>no order</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>23698</th>\n",
       "      <td>aa83af3a55f543d190400dc9392cb80f</td>\n",
       "      <td>M</td>\n",
       "      <td>76</td>\n",
       "      <td>54000.0</td>\n",
       "      <td>discount</td>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>0</td>\n",
       "      <td>no order</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>23873</th>\n",
       "      <td>aba68a20aa8a42acbe25ebcf0800414b</td>\n",
       "      <td>M</td>\n",
       "      <td>24</td>\n",
       "      <td>51000.0</td>\n",
       "      <td>discount</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>no order</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11824</th>\n",
       "      <td>558cb2d5299f4ddd99d684ccc3c8b189</td>\n",
       "      <td>M</td>\n",
       "      <td>63</td>\n",
       "      <td>55000.0</td>\n",
       "      <td>discount</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>no order</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                 person gender  age    income offer_type  \\\n",
       "15224  6ea759e2e1234242a48a575597924934      F   52   84000.0   discount   \n",
       "33807  f5112242df984403bfb4176d4487aef3      F   96   72000.0   discount   \n",
       "30111  da50429e4d43437eb8073e4264e686a5      M   61   82000.0       bogo   \n",
       "1209   093f0c09cc7242228edbb0e52ca14af6      M   36   35000.0   discount   \n",
       "26342  bd777dba9c784ff8aa96c2a4531fba1f      M   84   93000.0       bogo   \n",
       "25459  b702b01e1767462aa447431c9e8a07f9      M   57  110000.0       bogo   \n",
       "5533   2a070c1a63e348fda6f9772df48f4c85      M   73   36000.0   discount   \n",
       "23698  aa83af3a55f543d190400dc9392cb80f      M   76   54000.0   discount   \n",
       "23873  aba68a20aa8a42acbe25ebcf0800414b      M   24   51000.0   discount   \n",
       "11824  558cb2d5299f4ddd99d684ccc3c8b189      M   63   55000.0   discount   \n",
       "\n",
       "       event_offer_received  event_offer_viewed  event_offer_completed  \\\n",
       "15224                     2                   1                      1   \n",
       "33807                     3                   3                      2   \n",
       "30111                     1                   1                      1   \n",
       "1209                      3                   3                      3   \n",
       "26342                     3                   2                      2   \n",
       "25459                     1                   1                      1   \n",
       "5533                      1                   1                      0   \n",
       "23698                     2                   2                      0   \n",
       "23873                     1                   1                      0   \n",
       "11824                     1                   1                      0   \n",
       "\n",
       "                label  \n",
       "15224  possible order  \n",
       "33807  possible order  \n",
       "30111  possible order  \n",
       "1209   possible order  \n",
       "26342  possible order  \n",
       "25459  possible order  \n",
       "5533         no order  \n",
       "23698        no order  \n",
       "23873        no order  \n",
       "11824        no order  "
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# First, compare how many times the offer was viewed vs. received:\n",
    "df['label'] = np.where((df['event_offer_viewed'] >= df['event_offer_received']) &\n",
    "                       (df['event_offer_viewed'] > 0), \"no order\", \"no view\")\n",
    "\n",
    "# Next, compare how many times offer was completed vs. viewed\n",
    "# We use 0.5 threshold as in at least half of offers has to be completed:\n",
    "df['label'] = np.where(((df['event_offer_completed'] / df['event_offer_viewed']) >= 0.5) &\n",
    "                       (df['event_offer_viewed'] > 0) & \n",
    "                       (df['event_offer_completed'] > 0), \"possible order\", \"no order\")\n",
    "\n",
    "# Finally, if there is no order and no view:\n",
    "df['label'] = np.where((df['event_offer_viewed'] == 0) & \n",
    "                       (df['event_offer_completed'] == 0), \"no view\", df['label'])\n",
    "\n",
    "df.sample(10)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Drop the event columns, we don't need them anymore:\n",
    "df = df.drop(['event_offer_received', 'event_offer_viewed', 'event_offer_completed'], axis=1)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Convert labels to numeric"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "MultiIndex(levels=[['no order', 'no view', 'possible order'], [0.0, 1.0, 2.0]],\n",
       "           codes=[[0, 1, 2], [1, 0, 2]],\n",
       "           names=['label', 'output'])"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df[\"output\"] = np.where(df['label']=='no view', 0, np.nan)\n",
    "df[\"output\"] = np.where(df['label']=='no order', 1, df[\"output\"])\n",
    "df[\"output\"] = np.where(df['label']=='possible order', 2, df[\"output\"])\n",
    "\n",
    "df.groupby(['label', 'output']).mean().index"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Check for missing values"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "person        0\n",
       "gender        0\n",
       "age           0\n",
       "income        0\n",
       "offer_type    0\n",
       "label         0\n",
       "output        0\n",
       "dtype: int64"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.isnull().sum()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Encode \"gender\" and \"offer_type\" using a one-hot encoding scheme"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [],
   "source": [
    "df = df.join(pd.get_dummies(df['gender'])).drop('gender', axis=1) # drop the original column\n",
    "\n",
    "for column in ['M', 'F', 'O', 'U']:\n",
    "    df = df.rename(columns={column: (\"gender_\" + column.replace(\" \", \"_\"))}) # rename new columns\n",
    "    \n",
    "    \n",
    "df = df.join(pd.get_dummies(df['offer_type'])).drop('offer_type', axis=1) # drop the original column"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Drop the \"person\" and \"label\" columns"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>age</th>\n",
       "      <th>income</th>\n",
       "      <th>output</th>\n",
       "      <th>gender_F</th>\n",
       "      <th>gender_M</th>\n",
       "      <th>gender_O</th>\n",
       "      <th>bogo</th>\n",
       "      <th>discount</th>\n",
       "      <th>informational</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>25690</th>\n",
       "      <td>44</td>\n",
       "      <td>68000.0</td>\n",
       "      <td>2.0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1079</th>\n",
       "      <td>58</td>\n",
       "      <td>55000.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3955</th>\n",
       "      <td>48</td>\n",
       "      <td>71000.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "       age   income  output  gender_F  gender_M  gender_O  bogo  discount  \\\n",
       "25690   44  68000.0     2.0         0         1         0     0         1   \n",
       "1079    58  55000.0     1.0         0         1         0     0         1   \n",
       "3955    48  71000.0     0.0         1         0         0     0         1   \n",
       "\n",
       "       informational  \n",
       "25690              0  \n",
       "1079               0  \n",
       "3955               0  "
      ]
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df = df.drop(['person', 'label'], axis=1)\n",
    "\n",
    "df.sample(3)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(35324, 9)"
      ]
     },
     "execution_count": 14,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.shape"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Split the data to input features (X) and output labels (y)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "X:\n",
      "   age   income  gender_F  gender_M  gender_O  bogo  discount  informational\n",
      "0   33  72000.0         0         1         0     1         0              0\n",
      "1   33  72000.0         0         1         0     0         1              0\n",
      "2   33  72000.0         0         1         0     0         0              1\n",
      "3   40  57000.0         0         0         1     1         0              0\n",
      "4   40  57000.0         0         0         1     0         1              0\n",
      "\n",
      "\n",
      "\n",
      "y:\n",
      "0    1.0\n",
      "1    2.0\n",
      "2    2.0\n",
      "3    2.0\n",
      "4    2.0\n",
      "Name: output, dtype: float64\n"
     ]
    }
   ],
   "source": [
    "X_df = df.drop(['output'], axis=1)\n",
    "y_df = df['output']\n",
    "\n",
    "print(\"X:\")\n",
    "print(X_df.head())\n",
    "\n",
    "print(\"\\n\\n\\ny:\")\n",
    "print(y_df.head())"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Split the rows up into train, test and validation sets"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "X training set:\n",
      "(15856, 8)\n",
      "\n",
      "X valitation set:\n",
      "(7811, 8)\n",
      "\n",
      "X test set:\n",
      "(11657, 8)\n"
     ]
    }
   ],
   "source": [
    "# Split the dataset into 2/3 training and 1/3 testing sets:\n",
    "X_train, X_test, y_train, y_test = sklearn.model_selection.train_test_split(X_df, y_df, test_size=0.33)\n",
    "\n",
    "# Next, split the TRAINING set further into 2/3 training and 1/3 validation sets:\n",
    "X_train, X_val, y_train, y_val = sklearn.model_selection.train_test_split(X_train, y_train, test_size=0.33)\n",
    "\n",
    "print(\"X training set:\")\n",
    "print(X_train.shape)\n",
    "\n",
    "print(\"\\nX valitation set:\")\n",
    "print(X_val.shape)   \n",
    "                                                                          \n",
    "print(\"\\nX test set:\")\n",
    "print(X_test.shape)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Save the data locally as CSV files\n",
    "\n",
    "Amazon SageMaker requires that a CSV file **doesn't have a header** record and that the **target variable is in the first column** (from https://docs.aws.amazon.com/sagemaker/latest/dg/cdf-training.html)."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [],
   "source": [
    "data_dir = '../data' \n",
    "if not os.path.exists(data_dir):\n",
    "    os.makedirs(data_dir)\n",
    "\n",
    "# Save testing set:\n",
    "X_test.to_csv(os.path.join(data_dir, 'test.csv'), header=False, index=False)\n",
    "\n",
    "# Save training and validation sets:\n",
    "pd.concat([y_val, X_val], axis=1).to_csv(os.path.join(data_dir, 'validation.csv'), header=False, index=False)\n",
    "pd.concat([y_train, X_train], axis=1).to_csv(os.path.join(data_dir, 'train.csv'), header=False, index=False)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Upload the data files to S3"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [],
   "source": [
    "prefix = 'starbucks-xgboost'\n",
    "\n",
    "session = sagemaker.Session()\n",
    "\n",
    "test_location = session.upload_data(os.path.join(data_dir, 'test.csv'), key_prefix=prefix)\n",
    "val_location = session.upload_data(os.path.join(data_dir, 'validation.csv'), key_prefix=prefix)\n",
    "train_location = session.upload_data(os.path.join(data_dir, 'train.csv'), key_prefix=prefix)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Training the XGBoost model"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Construct the estimator object"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [],
   "source": [
    "role = get_execution_role()\n",
    "\n",
    "# Get the URI for new container:\n",
    "container = get_image_uri(session.boto_region_name, 'xgboost', '0.90-1')\n",
    "\n",
    "# Construct the estimator object:\n",
    "xgb = sagemaker.estimator.Estimator(container,\n",
    "                                    role,\n",
    "                                    train_instance_count=1,\n",
    "                                    train_instance_type='ml.m4.xlarge',\n",
    "                                    output_path='s3://{}/{}/output'.format(session.default_bucket(), prefix),\n",
    "                                    sagemaker_session=session) "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Set model specific hyperparameter default values\n",
    "\n",
    "We will use \"**multi:softprob**\" objective, which outputs a vector of ndata * nclass, which can be further reshaped to ndata * nclass matrix. The result contains predicted probability of each data point belonging to each class (from https://github.com/dmlc/xgboost/blob/master/doc/parameter.rst)."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [],
   "source": [
    "xgb.set_hyperparameters(num_round=500, # the number of rounds to run the training\n",
    "                        num_class=3, # the number of output classes\n",
    "                        objective='multi:softprob')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Train model on default parameters"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "2020-02-01 16:54:47 Starting - Starting the training job...\n",
      "2020-02-01 16:54:49 Starting - Launching requested ML instances......\n",
      "2020-02-01 16:55:53 Starting - Preparing the instances for training......\n",
      "2020-02-01 16:57:14 Downloading - Downloading input data\n",
      "2020-02-01 16:57:14 Training - Downloading the training image...\n",
      "2020-02-01 16:57:45 Training - Training image download completed. Training in progress...\u001b[34mINFO:sagemaker-containers:Imported framework sagemaker_xgboost_container.training\u001b[0m\n",
      "\u001b[34mINFO:sagemaker-containers:Failed to parse hyperparameter objective value multi:softprob to Json.\u001b[0m\n",
      "\u001b[34mReturning the value itself\u001b[0m\n",
      "\u001b[34mINFO:sagemaker-containers:No GPUs detected (normal if no gpus installed)\u001b[0m\n",
      "\u001b[34mINFO:sagemaker_xgboost_container.training:Running XGBoost Sagemaker in algorithm mode\u001b[0m\n",
      "\u001b[34mINFO:root:Determined delimiter of CSV input is ','\u001b[0m\n",
      "\u001b[34mINFO:root:Determined delimiter of CSV input is ','\u001b[0m\n",
      "\u001b[34mINFO:root:Determined delimiter of CSV input is ','\u001b[0m\n",
      "\u001b[34m[16:57:48] 15856x8 matrix with 126848 entries loaded from /opt/ml/input/data/train?format=csv&label_column=0&delimiter=,\u001b[0m\n",
      "\u001b[34mINFO:root:Determined delimiter of CSV input is ','\u001b[0m\n",
      "\u001b[34m[16:57:48] 7811x8 matrix with 62488 entries loaded from /opt/ml/input/data/validation?format=csv&label_column=0&delimiter=,\u001b[0m\n",
      "\u001b[34mINFO:root:Single node training.\u001b[0m\n",
      "\u001b[34mINFO:root:Train matrix has 15856 rows\u001b[0m\n",
      "\u001b[34mINFO:root:Validation matrix has 7811 rows\u001b[0m\n",
      "\u001b[34m[0]#011train-merror:0.390578#011validation-merror:0.397772\u001b[0m\n",
      "\u001b[34m[1]#011train-merror:0.389506#011validation-merror:0.39726\u001b[0m\n",
      "\u001b[34m[2]#011train-merror:0.389695#011validation-merror:0.397132\u001b[0m\n",
      "\u001b[34m[3]#011train-merror:0.388686#011validation-merror:0.39662\u001b[0m\n",
      "\u001b[34m[4]#011train-merror:0.388749#011validation-merror:0.395852\u001b[0m\n",
      "\u001b[34m[5]#011train-merror:0.38774#011validation-merror:0.396364\u001b[0m\n",
      "\u001b[34m[6]#011train-merror:0.38755#011validation-merror:0.39726\u001b[0m\n",
      "\u001b[34m[7]#011train-merror:0.387046#011validation-merror:0.39662\u001b[0m\n",
      "\u001b[34m[8]#011train-merror:0.3861#011validation-merror:0.396748\u001b[0m\n",
      "\u001b[34m[9]#011train-merror:0.386226#011validation-merror:0.397004\u001b[0m\n",
      "\u001b[34m[10]#011train-merror:0.385785#011validation-merror:0.396364\u001b[0m\n",
      "\u001b[34m[11]#011train-merror:0.385091#011validation-merror:0.39662\u001b[0m\n",
      "\u001b[34m[12]#011train-merror:0.38446#011validation-merror:0.39726\u001b[0m\n",
      "\u001b[34m[13]#011train-merror:0.384208#011validation-merror:0.397004\u001b[0m\n",
      "\u001b[34m[14]#011train-merror:0.383829#011validation-merror:0.398028\u001b[0m\n",
      "\u001b[34m[15]#011train-merror:0.38301#011validation-merror:0.3979\u001b[0m\n",
      "\u001b[34m[16]#011train-merror:0.382253#011validation-merror:0.398028\u001b[0m\n",
      "\u001b[34m[17]#011train-merror:0.381811#011validation-merror:0.398028\u001b[0m\n",
      "\u001b[34m[18]#011train-merror:0.381054#011validation-merror:0.398284\u001b[0m\n",
      "\u001b[34m[19]#011train-merror:0.380235#011validation-merror:0.399309\u001b[0m\n",
      "\u001b[34m[20]#011train-merror:0.379415#011validation-merror:0.399565\u001b[0m\n",
      "\u001b[34m[21]#011train-merror:0.379226#011validation-merror:0.400205\u001b[0m\n",
      "\u001b[34m[22]#011train-merror:0.378153#011validation-merror:0.399309\u001b[0m\n",
      "\u001b[34m[23]#011train-merror:0.376766#011validation-merror:0.400333\u001b[0m\n",
      "\u001b[34m[24]#011train-merror:0.375883#011validation-merror:0.399821\u001b[0m\n",
      "\u001b[34m[25]#011train-merror:0.375757#011validation-merror:0.399565\u001b[0m\n",
      "\u001b[34m[26]#011train-merror:0.374874#011validation-merror:0.399181\u001b[0m\n",
      "\u001b[34m[27]#011train-merror:0.373991#011validation-merror:0.399309\u001b[0m\n",
      "\u001b[34m[28]#011train-merror:0.37336#011validation-merror:0.399309\u001b[0m\n",
      "\u001b[34m[29]#011train-merror:0.372919#011validation-merror:0.399053\u001b[0m\n",
      "\u001b[34m[30]#011train-merror:0.372477#011validation-merror:0.399437\u001b[0m\n",
      "\u001b[34m[31]#011train-merror:0.371531#011validation-merror:0.399309\u001b[0m\n",
      "\u001b[34m[32]#011train-merror:0.371153#011validation-merror:0.399053\u001b[0m\n",
      "\u001b[34m[33]#011train-merror:0.370774#011validation-merror:0.400333\u001b[0m\n",
      "\u001b[34m[34]#011train-merror:0.370207#011validation-merror:0.400077\u001b[0m\n",
      "\u001b[34m[35]#011train-merror:0.370018#011validation-merror:0.400077\u001b[0m\n",
      "\u001b[34m[36]#011train-merror:0.369576#011validation-merror:0.400333\u001b[0m\n",
      "\u001b[34m[37]#011train-merror:0.369261#011validation-merror:0.400973\u001b[0m\n",
      "\u001b[34m[38]#011train-merror:0.368252#011validation-merror:0.401741\u001b[0m\n",
      "\u001b[34m[39]#011train-merror:0.366738#011validation-merror:0.402509\u001b[0m\n",
      "\u001b[34m[40]#011train-merror:0.366549#011validation-merror:0.402893\u001b[0m\n",
      "\u001b[34m[41]#011train-merror:0.366107#011validation-merror:0.402637\u001b[0m\n",
      "\u001b[34m[42]#011train-merror:0.364972#011validation-merror:0.403918\u001b[0m\n",
      "\u001b[34m[43]#011train-merror:0.364468#011validation-merror:0.403405\u001b[0m\n",
      "\u001b[34m[44]#011train-merror:0.3639#011validation-merror:0.403533\u001b[0m\n",
      "\u001b[34m[45]#011train-merror:0.363396#011validation-merror:0.403405\u001b[0m\n",
      "\u001b[34m[46]#011train-merror:0.36308#011validation-merror:0.404046\u001b[0m\n",
      "\u001b[34m[47]#011train-merror:0.36308#011validation-merror:0.40379\u001b[0m\n",
      "\u001b[34m[48]#011train-merror:0.361693#011validation-merror:0.404686\u001b[0m\n",
      "\u001b[34m[49]#011train-merror:0.35999#011validation-merror:0.404686\u001b[0m\n",
      "\u001b[34m[50]#011train-merror:0.359612#011validation-merror:0.405582\u001b[0m\n",
      "\u001b[34m[51]#011train-merror:0.357278#011validation-merror:0.406478\u001b[0m\n",
      "\u001b[34m[52]#011train-merror:0.357341#011validation-merror:0.406478\u001b[0m\n",
      "\u001b[34m[53]#011train-merror:0.355827#011validation-merror:0.40635\u001b[0m\n",
      "\u001b[34m[54]#011train-merror:0.355386#011validation-merror:0.406094\u001b[0m\n",
      "\u001b[34m[55]#011train-merror:0.354881#011validation-merror:0.40699\u001b[0m\n",
      "\u001b[34m[56]#011train-merror:0.354125#011validation-merror:0.406094\u001b[0m\n",
      "\u001b[34m[57]#011train-merror:0.353809#011validation-merror:0.40699\u001b[0m\n",
      "\u001b[34m[58]#011train-merror:0.353305#011validation-merror:0.407886\u001b[0m\n",
      "\u001b[34m[59]#011train-merror:0.352737#011validation-merror:0.408526\u001b[0m\n",
      "\u001b[34m[60]#011train-merror:0.352106#011validation-merror:0.408526\u001b[0m\n",
      "\u001b[34m[61]#011train-merror:0.351728#011validation-merror:0.408526\u001b[0m\n",
      "\u001b[34m[62]#011train-merror:0.35135#011validation-merror:0.409167\u001b[0m\n",
      "\u001b[34m[63]#011train-merror:0.350277#011validation-merror:0.409551\u001b[0m\n",
      "\u001b[34m[64]#011train-merror:0.349773#011validation-merror:0.410319\u001b[0m\n",
      "\u001b[34m[65]#011train-merror:0.349395#011validation-merror:0.410191\u001b[0m\n",
      "\u001b[34m[66]#011train-merror:0.349268#011validation-merror:0.410191\u001b[0m\n",
      "\u001b[34m[67]#011train-merror:0.349142#011validation-merror:0.410319\u001b[0m\n",
      "\u001b[34m[68]#011train-merror:0.349142#011validation-merror:0.410063\u001b[0m\n",
      "\u001b[34m[69]#011train-merror:0.348512#011validation-merror:0.410063\u001b[0m\n",
      "\u001b[34m[70]#011train-merror:0.347566#011validation-merror:0.410575\u001b[0m\n",
      "\u001b[34m[71]#011train-merror:0.347061#011validation-merror:0.411215\u001b[0m\n",
      "\u001b[34m[72]#011train-merror:0.346304#011validation-merror:0.412239\u001b[0m\n",
      "\u001b[34m[73]#011train-merror:0.345989#011validation-merror:0.412495\u001b[0m\n",
      "\u001b[34m[74]#011train-merror:0.344854#011validation-merror:0.412751\u001b[0m\n",
      "\u001b[34m[75]#011train-merror:0.344917#011validation-merror:0.412495\u001b[0m\n",
      "\u001b[34m[76]#011train-merror:0.344349#011validation-merror:0.413135\u001b[0m\n",
      "\u001b[34m[77]#011train-merror:0.343908#011validation-merror:0.412239\u001b[0m\n",
      "\u001b[34m[78]#011train-merror:0.34334#011validation-merror:0.413135\u001b[0m\n",
      "\u001b[34m[79]#011train-merror:0.342836#011validation-merror:0.413007\u001b[0m\n",
      "\u001b[34m[80]#011train-merror:0.342962#011validation-merror:0.412879\u001b[0m\n",
      "\u001b[34m[81]#011train-merror:0.341763#011validation-merror:0.413263\u001b[0m\n",
      "\u001b[34m[82]#011train-merror:0.341259#011validation-merror:0.413775\u001b[0m\n",
      "\u001b[34m[83]#011train-merror:0.340754#011validation-merror:0.413903\u001b[0m\n",
      "\u001b[34m[84]#011train-merror:0.339619#011validation-merror:0.414032\u001b[0m\n",
      "\u001b[34m[85]#011train-merror:0.339367#011validation-merror:0.414416\u001b[0m\n",
      "\u001b[34m[86]#011train-merror:0.338925#011validation-merror:0.414416\u001b[0m\n",
      "\u001b[34m[87]#011train-merror:0.338799#011validation-merror:0.414928\u001b[0m\n",
      "\u001b[34m[88]#011train-merror:0.33861#011validation-merror:0.41416\u001b[0m\n",
      "\u001b[34m[89]#011train-merror:0.337979#011validation-merror:0.41416\u001b[0m\n",
      "\u001b[34m[90]#011train-merror:0.33779#011validation-merror:0.413903\u001b[0m\n",
      "\u001b[34m[91]#011train-merror:0.337601#011validation-merror:0.4148\u001b[0m\n",
      "\u001b[34m[92]#011train-merror:0.337601#011validation-merror:0.414928\u001b[0m\n",
      "\u001b[34m[93]#011train-merror:0.336213#011validation-merror:0.4148\u001b[0m\n",
      "\u001b[34m[94]#011train-merror:0.33615#011validation-merror:0.415184\u001b[0m\n",
      "\u001b[34m[95]#011train-merror:0.335204#011validation-merror:0.41608\u001b[0m\n",
      "\u001b[34m[96]#011train-merror:0.334889#011validation-merror:0.41608\u001b[0m\n",
      "\u001b[34m[97]#011train-merror:0.334511#011validation-merror:0.41608\u001b[0m\n",
      "\u001b[34m[98]#011train-merror:0.334321#011validation-merror:0.416208\u001b[0m\n",
      "\u001b[34m[99]#011train-merror:0.334006#011validation-merror:0.41608\u001b[0m\n",
      "\u001b[34m[100]#011train-merror:0.333249#011validation-merror:0.415824\u001b[0m\n",
      "\u001b[34m[101]#011train-merror:0.332682#011validation-merror:0.415952\u001b[0m\n",
      "\u001b[34m[102]#011train-merror:0.331609#011validation-merror:0.41544\u001b[0m\n",
      "\u001b[34m[103]#011train-merror:0.331609#011validation-merror:0.414928\u001b[0m\n",
      "\u001b[34m[104]#011train-merror:0.331357#011validation-merror:0.415184\u001b[0m\n",
      "\u001b[34m[105]#011train-merror:0.331105#011validation-merror:0.415056\u001b[0m\n",
      "\u001b[34m[106]#011train-merror:0.330663#011validation-merror:0.4148\u001b[0m\n",
      "\u001b[34m[107]#011train-merror:0.330348#011validation-merror:0.41544\u001b[0m\n",
      "\u001b[34m[108]#011train-merror:0.330159#011validation-merror:0.415056\u001b[0m\n",
      "\u001b[34m[109]#011train-merror:0.329654#011validation-merror:0.4148\u001b[0m\n",
      "\u001b[34m[110]#011train-merror:0.328645#011validation-merror:0.415184\u001b[0m\n",
      "\u001b[34m[111]#011train-merror:0.32833#011validation-merror:0.414416\u001b[0m\n",
      "\u001b[34m[112]#011train-merror:0.32751#011validation-merror:0.414416\u001b[0m\n",
      "\u001b[34m[113]#011train-merror:0.327132#011validation-merror:0.4148\u001b[0m\n",
      "\u001b[34m[114]#011train-merror:0.327573#011validation-merror:0.415312\u001b[0m\n",
      "\u001b[34m[115]#011train-merror:0.327636#011validation-merror:0.415568\u001b[0m\n",
      "\u001b[34m[116]#011train-merror:0.326816#011validation-merror:0.415824\u001b[0m\n",
      "\u001b[34m[117]#011train-merror:0.326438#011validation-merror:0.415568\u001b[0m\n",
      "\u001b[34m[118]#011train-merror:0.325681#011validation-merror:0.417488\u001b[0m\n",
      "\u001b[34m[119]#011train-merror:0.32505#011validation-merror:0.417232\u001b[0m\n",
      "\u001b[34m[120]#011train-merror:0.324104#011validation-merror:0.417872\u001b[0m\n",
      "\u001b[34m[121]#011train-merror:0.324168#011validation-merror:0.418\u001b[0m\n",
      "\u001b[34m[122]#011train-merror:0.323537#011validation-merror:0.418256\u001b[0m\n",
      "\u001b[34m[123]#011train-merror:0.323222#011validation-merror:0.418768\u001b[0m\n",
      "\u001b[34m[124]#011train-merror:0.322843#011validation-merror:0.41864\u001b[0m\n",
      "\u001b[34m[125]#011train-merror:0.322275#011validation-merror:0.417872\u001b[0m\n",
      "\u001b[34m[126]#011train-merror:0.322212#011validation-merror:0.417488\u001b[0m\n",
      "\u001b[34m[127]#011train-merror:0.321645#011validation-merror:0.418384\u001b[0m\n",
      "\u001b[34m[128]#011train-merror:0.321393#011validation-merror:0.417744\u001b[0m\n",
      "\u001b[34m[129]#011train-merror:0.321014#011validation-merror:0.418\u001b[0m\n",
      "\u001b[34m[130]#011train-merror:0.320636#011validation-merror:0.418\u001b[0m\n",
      "\u001b[34m[131]#011train-merror:0.320257#011validation-merror:0.417872\u001b[0m\n",
      "\u001b[34m[132]#011train-merror:0.320257#011validation-merror:0.417872\u001b[0m\n",
      "\u001b[34m[133]#011train-merror:0.319816#011validation-merror:0.418256\u001b[0m\n",
      "\u001b[34m[134]#011train-merror:0.319311#011validation-merror:0.418384\u001b[0m\n",
      "\u001b[34m[135]#011train-merror:0.318681#011validation-merror:0.419152\u001b[0m\n",
      "\u001b[34m[136]#011train-merror:0.318239#011validation-merror:0.418896\u001b[0m\n",
      "\u001b[34m[137]#011train-merror:0.317419#011validation-merror:0.41864\u001b[0m\n",
      "\u001b[34m[138]#011train-merror:0.317482#011validation-merror:0.418384\u001b[0m\n",
      "\u001b[34m[139]#011train-merror:0.316978#011validation-merror:0.41864\u001b[0m\n",
      "\u001b[34m[140]#011train-merror:0.315843#011validation-merror:0.418896\u001b[0m\n",
      "\u001b[34m[141]#011train-merror:0.31496#011validation-merror:0.41928\u001b[0m\n",
      "\u001b[34m[142]#011train-merror:0.314392#011validation-merror:0.419921\u001b[0m\n",
      "\u001b[34m[143]#011train-merror:0.313446#011validation-merror:0.420817\u001b[0m\n",
      "\u001b[34m[144]#011train-merror:0.312878#011validation-merror:0.420561\u001b[0m\n",
      "\u001b[34m[145]#011train-merror:0.312941#011validation-merror:0.420945\u001b[0m\n",
      "\u001b[34m[146]#011train-merror:0.312878#011validation-merror:0.420945\u001b[0m\n",
      "\u001b[34m[147]#011train-merror:0.312626#011validation-merror:0.421073\u001b[0m\n",
      "\u001b[34m[148]#011train-merror:0.3125#011validation-merror:0.421457\u001b[0m\n",
      "\u001b[34m[149]#011train-merror:0.311932#011validation-merror:0.421457\u001b[0m\n",
      "\u001b[34m[150]#011train-merror:0.311932#011validation-merror:0.421201\u001b[0m\n",
      "\u001b[34m[151]#011train-merror:0.311932#011validation-merror:0.421073\u001b[0m\n",
      "\u001b[34m[152]#011train-merror:0.311491#011validation-merror:0.421201\u001b[0m\n",
      "\u001b[34m[153]#011train-merror:0.311302#011validation-merror:0.421585\u001b[0m\n",
      "\u001b[34m[154]#011train-merror:0.311239#011validation-merror:0.421201\u001b[0m\n",
      "\u001b[34m[155]#011train-merror:0.310419#011validation-merror:0.421329\u001b[0m\n",
      "\u001b[34m[156]#011train-merror:0.310167#011validation-merror:0.421585\u001b[0m\n",
      "\u001b[34m[157]#011train-merror:0.309977#011validation-merror:0.421585\u001b[0m\n",
      "\u001b[34m[158]#011train-merror:0.310167#011validation-merror:0.421713\u001b[0m\n",
      "\u001b[34m[159]#011train-merror:0.31023#011validation-merror:0.421329\u001b[0m\n",
      "\u001b[34m[160]#011train-merror:0.310293#011validation-merror:0.421201\u001b[0m\n",
      "\u001b[34m[161]#011train-merror:0.309662#011validation-merror:0.421841\u001b[0m\n",
      "\u001b[34m[162]#011train-merror:0.308842#011validation-merror:0.422353\u001b[0m\n",
      "\u001b[34m[163]#011train-merror:0.308779#011validation-merror:0.422737\u001b[0m\n",
      "\u001b[34m[164]#011train-merror:0.308401#011validation-merror:0.423377\u001b[0m\n",
      "\u001b[34m[165]#011train-merror:0.307896#011validation-merror:0.423761\u001b[0m\n",
      "\u001b[34m[166]#011train-merror:0.307896#011validation-merror:0.424017\u001b[0m\n",
      "\u001b[34m[167]#011train-merror:0.307076#011validation-merror:0.423505\u001b[0m\n",
      "\u001b[34m[168]#011train-merror:0.306572#011validation-merror:0.423249\u001b[0m\n",
      "\u001b[34m[169]#011train-merror:0.306193#011validation-merror:0.423633\u001b[0m\n",
      "\u001b[34m[170]#011train-merror:0.306256#011validation-merror:0.424017\u001b[0m\n",
      "\u001b[34m[171]#011train-merror:0.305436#011validation-merror:0.423249\u001b[0m\n",
      "\u001b[34m[172]#011train-merror:0.304995#011validation-merror:0.423377\u001b[0m\n",
      "\u001b[34m[173]#011train-merror:0.304806#011validation-merror:0.423761\u001b[0m\n",
      "\u001b[34m[174]#011train-merror:0.304238#011validation-merror:0.424017\u001b[0m\n",
      "\u001b[34m[175]#011train-merror:0.303481#011validation-merror:0.423633\u001b[0m\n",
      "\u001b[34m[176]#011train-merror:0.30304#011validation-merror:0.423889\u001b[0m\n",
      "\u001b[34m[177]#011train-merror:0.302725#011validation-merror:0.424017\u001b[0m\n",
      "\u001b[34m[178]#011train-merror:0.301715#011validation-merror:0.423121\u001b[0m\n",
      "\u001b[34m[179]#011train-merror:0.301526#011validation-merror:0.423121\u001b[0m\n",
      "\u001b[34m[180]#011train-merror:0.301778#011validation-merror:0.422737\u001b[0m\n",
      "\u001b[34m[181]#011train-merror:0.301211#011validation-merror:0.422609\u001b[0m\n",
      "\u001b[34m[182]#011train-merror:0.301022#011validation-merror:0.422865\u001b[0m\n",
      "\u001b[34m[183]#011train-merror:0.301022#011validation-merror:0.422737\u001b[0m\n",
      "\u001b[34m[184]#011train-merror:0.300769#011validation-merror:0.422865\u001b[0m\n",
      "\u001b[34m[185]#011train-merror:0.300706#011validation-merror:0.423121\u001b[0m\n",
      "\u001b[34m[186]#011train-merror:0.300517#011validation-merror:0.423761\u001b[0m\n",
      "\u001b[34m[187]#011train-merror:0.300328#011validation-merror:0.423505\u001b[0m\n",
      "\u001b[34m[188]#011train-merror:0.300391#011validation-merror:0.423761\u001b[0m\n",
      "\u001b[34m[189]#011train-merror:0.300454#011validation-merror:0.423633\u001b[0m\n",
      "\u001b[34m[190]#011train-merror:0.300013#011validation-merror:0.423505\u001b[0m\n",
      "\u001b[34m[191]#011train-merror:0.299193#011validation-merror:0.422737\u001b[0m\n",
      "\u001b[34m[192]#011train-merror:0.298247#011validation-merror:0.423377\u001b[0m\n",
      "\u001b[34m[193]#011train-merror:0.298184#011validation-merror:0.422737\u001b[0m\n",
      "\u001b[34m[194]#011train-merror:0.298247#011validation-merror:0.423505\u001b[0m\n",
      "\u001b[34m[195]#011train-merror:0.297931#011validation-merror:0.424145\u001b[0m\n",
      "\u001b[34m[196]#011train-merror:0.297301#011validation-merror:0.424401\u001b[0m\n",
      "\u001b[34m[197]#011train-merror:0.297553#011validation-merror:0.424273\u001b[0m\n",
      "\u001b[34m[198]#011train-merror:0.29749#011validation-merror:0.424145\u001b[0m\n",
      "\u001b[34m[199]#011train-merror:0.297175#011validation-merror:0.42517\u001b[0m\n",
      "\u001b[34m[200]#011train-merror:0.296292#011validation-merror:0.425298\u001b[0m\n",
      "\u001b[34m[201]#011train-merror:0.296292#011validation-merror:0.425426\u001b[0m\n",
      "\u001b[34m[202]#011train-merror:0.296165#011validation-merror:0.425298\u001b[0m\n",
      "\u001b[34m[203]#011train-merror:0.296039#011validation-merror:0.42517\u001b[0m\n",
      "\u001b[34m[204]#011train-merror:0.296165#011validation-merror:0.425554\u001b[0m\n",
      "\u001b[34m[205]#011train-merror:0.296039#011validation-merror:0.426194\u001b[0m\n",
      "\u001b[34m[206]#011train-merror:0.295724#011validation-merror:0.426194\u001b[0m\n",
      "\u001b[34m[207]#011train-merror:0.295535#011validation-merror:0.426322\u001b[0m\n",
      "\u001b[34m[208]#011train-merror:0.294967#011validation-merror:0.427602\u001b[0m\n",
      "\u001b[34m[209]#011train-merror:0.294652#011validation-merror:0.42773\u001b[0m\n",
      "\u001b[34m[210]#011train-merror:0.294337#011validation-merror:0.427346\u001b[0m\n",
      "\u001b[34m[211]#011train-merror:0.294273#011validation-merror:0.427346\u001b[0m\n",
      "\u001b[34m[212]#011train-merror:0.294337#011validation-merror:0.427218\u001b[0m\n",
      "\u001b[34m[213]#011train-merror:0.294147#011validation-merror:0.42773\u001b[0m\n",
      "\u001b[34m[214]#011train-merror:0.293895#011validation-merror:0.428114\u001b[0m\n",
      "\u001b[34m[215]#011train-merror:0.293643#011validation-merror:0.42773\u001b[0m\n",
      "\u001b[34m[216]#011train-merror:0.293327#011validation-merror:0.427858\u001b[0m\n",
      "\u001b[34m[217]#011train-merror:0.293075#011validation-merror:0.427858\u001b[0m\n",
      "\u001b[34m[218]#011train-merror:0.29276#011validation-merror:0.42773\u001b[0m\n",
      "\u001b[34m[219]#011train-merror:0.292192#011validation-merror:0.427986\u001b[0m\n",
      "\u001b[34m[220]#011train-merror:0.291751#011validation-merror:0.427602\u001b[0m\n",
      "\u001b[34m[221]#011train-merror:0.291435#011validation-merror:0.427858\u001b[0m\n",
      "\u001b[34m[222]#011train-merror:0.290868#011validation-merror:0.427858\u001b[0m\n",
      "\u001b[34m[223]#011train-merror:0.290805#011validation-merror:0.427474\u001b[0m\n",
      "\u001b[34m[224]#011train-merror:0.290237#011validation-merror:0.427858\u001b[0m\n",
      "\u001b[34m[225]#011train-merror:0.289859#011validation-merror:0.428242\u001b[0m\n",
      "\u001b[34m[226]#011train-merror:0.289796#011validation-merror:0.428114\u001b[0m\n",
      "\u001b[34m[227]#011train-merror:0.289606#011validation-merror:0.427858\u001b[0m\n",
      "\u001b[34m[228]#011train-merror:0.28967#011validation-merror:0.42773\u001b[0m\n",
      "\u001b[34m[229]#011train-merror:0.289228#011validation-merror:0.42773\u001b[0m\n",
      "\u001b[34m[230]#011train-merror:0.289039#011validation-merror:0.427474\u001b[0m\n",
      "\u001b[34m[231]#011train-merror:0.288219#011validation-merror:0.427346\u001b[0m\n",
      "\u001b[34m[232]#011train-merror:0.287904#011validation-merror:0.426834\u001b[0m\n",
      "\u001b[34m[233]#011train-merror:0.287841#011validation-merror:0.427858\u001b[0m\n",
      "\u001b[34m[234]#011train-merror:0.287651#011validation-merror:0.428242\u001b[0m\n",
      "\u001b[34m[235]#011train-merror:0.287399#011validation-merror:0.429266\u001b[0m\n",
      "\u001b[34m[236]#011train-merror:0.287147#011validation-merror:0.428754\u001b[0m\n",
      "\u001b[34m[237]#011train-merror:0.287084#011validation-merror:0.428754\u001b[0m\n",
      "\u001b[34m[238]#011train-merror:0.287273#011validation-merror:0.428626\u001b[0m\n",
      "\u001b[34m[239]#011train-merror:0.287021#011validation-merror:0.428626\u001b[0m\n",
      "\u001b[34m[240]#011train-merror:0.286768#011validation-merror:0.42837\u001b[0m\n",
      "\u001b[34m[241]#011train-merror:0.286705#011validation-merror:0.430419\u001b[0m\n",
      "\u001b[34m[242]#011train-merror:0.286201#011validation-merror:0.429779\u001b[0m\n",
      "\u001b[34m[243]#011train-merror:0.285822#011validation-merror:0.428754\u001b[0m\n",
      "\u001b[34m[244]#011train-merror:0.285255#011validation-merror:0.428882\u001b[0m\n",
      "\u001b[34m[245]#011train-merror:0.284876#011validation-merror:0.429394\u001b[0m\n",
      "\u001b[34m[246]#011train-merror:0.284939#011validation-merror:0.429394\u001b[0m\n",
      "\u001b[34m[247]#011train-merror:0.285003#011validation-merror:0.429266\u001b[0m\n",
      "\u001b[34m[248]#011train-merror:0.284687#011validation-merror:0.429138\u001b[0m\n",
      "\u001b[34m[249]#011train-merror:0.284498#011validation-merror:0.429138\u001b[0m\n",
      "\u001b[34m[250]#011train-merror:0.284435#011validation-merror:0.42901\u001b[0m\n",
      "\u001b[34m[251]#011train-merror:0.283489#011validation-merror:0.429522\u001b[0m\n",
      "\u001b[34m[252]#011train-merror:0.283489#011validation-merror:0.429394\u001b[0m\n",
      "\u001b[34m[253]#011train-merror:0.2833#011validation-merror:0.42965\u001b[0m\n",
      "\u001b[34m[254]#011train-merror:0.282795#011validation-merror:0.430163\u001b[0m\n",
      "\u001b[34m[255]#011train-merror:0.282669#011validation-merror:0.430035\u001b[0m\n",
      "\u001b[34m[256]#011train-merror:0.282101#011validation-merror:0.430675\u001b[0m\n",
      "\u001b[34m[257]#011train-merror:0.282038#011validation-merror:0.430803\u001b[0m\n",
      "\u001b[34m[258]#011train-merror:0.281975#011validation-merror:0.430547\u001b[0m\n",
      "\u001b[34m[259]#011train-merror:0.281849#011validation-merror:0.430931\u001b[0m\n",
      "\u001b[34m[260]#011train-merror:0.281786#011validation-merror:0.430547\u001b[0m\n",
      "\u001b[34m[261]#011train-merror:0.281723#011validation-merror:0.430803\u001b[0m\n",
      "\u001b[34m[262]#011train-merror:0.28166#011validation-merror:0.430931\u001b[0m\n",
      "\u001b[34m[263]#011train-merror:0.281408#011validation-merror:0.431571\u001b[0m\n",
      "\u001b[34m[264]#011train-merror:0.281029#011validation-merror:0.432083\u001b[0m\n",
      "\u001b[34m[265]#011train-merror:0.280714#011validation-merror:0.432211\u001b[0m\n",
      "\u001b[34m[266]#011train-merror:0.280777#011validation-merror:0.432467\u001b[0m\n",
      "\u001b[34m[267]#011train-merror:0.280588#011validation-merror:0.432723\u001b[0m\n",
      "\u001b[34m[268]#011train-merror:0.280083#011validation-merror:0.433363\u001b[0m\n",
      "\u001b[34m[269]#011train-merror:0.279831#011validation-merror:0.433875\u001b[0m\n",
      "\u001b[34m[270]#011train-merror:0.279831#011validation-merror:0.434387\u001b[0m\n",
      "\u001b[34m[271]#011train-merror:0.27939#011validation-merror:0.434387\u001b[0m\n",
      "\u001b[34m[272]#011train-merror:0.278696#011validation-merror:0.434643\u001b[0m\n",
      "\u001b[34m[273]#011train-merror:0.27857#011validation-merror:0.434643\u001b[0m\n",
      "\u001b[34m[274]#011train-merror:0.278191#011validation-merror:0.434771\u001b[0m\n",
      "\u001b[34m[275]#011train-merror:0.27838#011validation-merror:0.435028\u001b[0m\n",
      "\u001b[34m[276]#011train-merror:0.27838#011validation-merror:0.435028\u001b[0m\n",
      "\u001b[34m[277]#011train-merror:0.278317#011validation-merror:0.434643\u001b[0m\n",
      "\u001b[34m[278]#011train-merror:0.277876#011validation-merror:0.434643\u001b[0m\n",
      "\u001b[34m[279]#011train-merror:0.277876#011validation-merror:0.434387\u001b[0m\n",
      "\u001b[34m[280]#011train-merror:0.278065#011validation-merror:0.43554\u001b[0m\n",
      "\u001b[34m[281]#011train-merror:0.278128#011validation-merror:0.435924\u001b[0m\n",
      "\u001b[34m[282]#011train-merror:0.278002#011validation-merror:0.436308\u001b[0m\n",
      "\u001b[34m[283]#011train-merror:0.277434#011validation-merror:0.436692\u001b[0m\n",
      "\u001b[34m[284]#011train-merror:0.276804#011validation-merror:0.437076\u001b[0m\n",
      "\u001b[34m[285]#011train-merror:0.276488#011validation-merror:0.436564\u001b[0m\n",
      "\u001b[34m[286]#011train-merror:0.27611#011validation-merror:0.436948\u001b[0m\n",
      "\u001b[34m[287]#011train-merror:0.275353#011validation-merror:0.43618\u001b[0m\n",
      "\u001b[34m[288]#011train-merror:0.275353#011validation-merror:0.435924\u001b[0m\n",
      "\u001b[34m[289]#011train-merror:0.275038#011validation-merror:0.436692\u001b[0m\n",
      "\u001b[34m[290]#011train-merror:0.275038#011validation-merror:0.437076\u001b[0m\n",
      "\u001b[34m[291]#011train-merror:0.275101#011validation-merror:0.436948\u001b[0m\n",
      "\u001b[34m[292]#011train-merror:0.274659#011validation-merror:0.436564\u001b[0m\n",
      "\u001b[34m[293]#011train-merror:0.274533#011validation-merror:0.43682\u001b[0m\n",
      "\u001b[34m[294]#011train-merror:0.274407#011validation-merror:0.437588\u001b[0m\n",
      "\u001b[34m[295]#011train-merror:0.274218#011validation-merror:0.437716\u001b[0m\n",
      "\u001b[34m[296]#011train-merror:0.274092#011validation-merror:0.437844\u001b[0m\n",
      "\u001b[34m[297]#011train-merror:0.273966#011validation-merror:0.437332\u001b[0m\n",
      "\u001b[34m[298]#011train-merror:0.273903#011validation-merror:0.436692\u001b[0m\n",
      "\u001b[34m[299]#011train-merror:0.27384#011validation-merror:0.43746\u001b[0m\n",
      "\u001b[34m[300]#011train-merror:0.273966#011validation-merror:0.437588\u001b[0m\n",
      "\u001b[34m[301]#011train-merror:0.273777#011validation-merror:0.437844\u001b[0m\n",
      "\u001b[34m[302]#011train-merror:0.273461#011validation-merror:0.4381\u001b[0m\n",
      "\u001b[34m[303]#011train-merror:0.273524#011validation-merror:0.437844\u001b[0m\n",
      "\u001b[34m[304]#011train-merror:0.273146#011validation-merror:0.437332\u001b[0m\n",
      "\u001b[34m[305]#011train-merror:0.272894#011validation-merror:0.437844\u001b[0m\n",
      "\u001b[34m[306]#011train-merror:0.271695#011validation-merror:0.438484\u001b[0m\n",
      "\u001b[34m[307]#011train-merror:0.271254#011validation-merror:0.43874\u001b[0m\n",
      "\u001b[34m[308]#011train-merror:0.271254#011validation-merror:0.43874\u001b[0m\n",
      "\u001b[34m[309]#011train-merror:0.271065#011validation-merror:0.438868\u001b[0m\n",
      "\u001b[34m[310]#011train-merror:0.270749#011validation-merror:0.438612\u001b[0m\n",
      "\u001b[34m[311]#011train-merror:0.270245#011validation-merror:0.438484\u001b[0m\n",
      "\u001b[34m[312]#011train-merror:0.270245#011validation-merror:0.438612\u001b[0m\n",
      "\u001b[34m[313]#011train-merror:0.270119#011validation-merror:0.438228\u001b[0m\n",
      "\u001b[34m[314]#011train-merror:0.270056#011validation-merror:0.438612\u001b[0m\n",
      "\u001b[34m[315]#011train-merror:0.269614#011validation-merror:0.43874\u001b[0m\n",
      "\u001b[34m[316]#011train-merror:0.269362#011validation-merror:0.438228\u001b[0m\n",
      "\u001b[34m[317]#011train-merror:0.268983#011validation-merror:0.438356\u001b[0m\n",
      "\u001b[34m[318]#011train-merror:0.269046#011validation-merror:0.438356\u001b[0m\n",
      "\u001b[34m[319]#011train-merror:0.268542#011validation-merror:0.438612\u001b[0m\n",
      "\u001b[34m[320]#011train-merror:0.268983#011validation-merror:0.438996\u001b[0m\n",
      "\u001b[34m[321]#011train-merror:0.268163#011validation-merror:0.438868\u001b[0m\n",
      "\u001b[34m[322]#011train-merror:0.268227#011validation-merror:0.439252\u001b[0m\n",
      "\u001b[34m[323]#011train-merror:0.267974#011validation-merror:0.439508\u001b[0m\n",
      "\u001b[34m[324]#011train-merror:0.267911#011validation-merror:0.438996\u001b[0m\n",
      "\u001b[34m[325]#011train-merror:0.267785#011validation-merror:0.439764\u001b[0m\n",
      "\u001b[34m[326]#011train-merror:0.267848#011validation-merror:0.439636\u001b[0m\n",
      "\u001b[34m[327]#011train-merror:0.267217#011validation-merror:0.440277\u001b[0m\n",
      "\u001b[34m[328]#011train-merror:0.267028#011validation-merror:0.440277\u001b[0m\n",
      "\u001b[34m[329]#011train-merror:0.267091#011validation-merror:0.440277\u001b[0m\n",
      "\u001b[34m[330]#011train-merror:0.266902#011validation-merror:0.440277\u001b[0m\n",
      "\u001b[34m[331]#011train-merror:0.266902#011validation-merror:0.440917\u001b[0m\n",
      "\u001b[34m[332]#011train-merror:0.266713#011validation-merror:0.440789\u001b[0m\n",
      "\u001b[34m[333]#011train-merror:0.266461#011validation-merror:0.440789\u001b[0m\n",
      "\u001b[34m[334]#011train-merror:0.266587#011validation-merror:0.440789\u001b[0m\n",
      "\u001b[34m[335]#011train-merror:0.266461#011validation-merror:0.441045\u001b[0m\n",
      "\u001b[34m[336]#011train-merror:0.266398#011validation-merror:0.441557\u001b[0m\n",
      "\u001b[34m[337]#011train-merror:0.265704#011validation-merror:0.441685\u001b[0m\n",
      "\u001b[34m[338]#011train-merror:0.265262#011validation-merror:0.442069\u001b[0m\n",
      "\u001b[34m[339]#011train-merror:0.264758#011validation-merror:0.442197\u001b[0m\n",
      "\u001b[34m[340]#011train-merror:0.26419#011validation-merror:0.442069\u001b[0m\n",
      "\u001b[34m[341]#011train-merror:0.264127#011validation-merror:0.442069\u001b[0m\n",
      "\u001b[34m[342]#011train-merror:0.264253#011validation-merror:0.442453\u001b[0m\n",
      "\u001b[34m[343]#011train-merror:0.26419#011validation-merror:0.442325\u001b[0m\n",
      "\u001b[34m[344]#011train-merror:0.263749#011validation-merror:0.442709\u001b[0m\n",
      "\u001b[34m[345]#011train-merror:0.263686#011validation-merror:0.442837\u001b[0m\n",
      "\u001b[34m[346]#011train-merror:0.263496#011validation-merror:0.442837\u001b[0m\n",
      "\u001b[34m[347]#011train-merror:0.263433#011validation-merror:0.442197\u001b[0m\n",
      "\u001b[34m[348]#011train-merror:0.263433#011validation-merror:0.441941\u001b[0m\n",
      "\u001b[34m[349]#011train-merror:0.263433#011validation-merror:0.441685\u001b[0m\n",
      "\u001b[34m[350]#011train-merror:0.263307#011validation-merror:0.441685\u001b[0m\n",
      "\u001b[34m[351]#011train-merror:0.263244#011validation-merror:0.441173\u001b[0m\n",
      "\u001b[34m[352]#011train-merror:0.263244#011validation-merror:0.442069\u001b[0m\n",
      "\u001b[34m[353]#011train-merror:0.262866#011validation-merror:0.442069\u001b[0m\n",
      "\u001b[34m[354]#011train-merror:0.262866#011validation-merror:0.442197\u001b[0m\n",
      "\u001b[34m[355]#011train-merror:0.26274#011validation-merror:0.441813\u001b[0m\n",
      "\u001b[34m[356]#011train-merror:0.26255#011validation-merror:0.441557\u001b[0m\n",
      "\u001b[34m[357]#011train-merror:0.262361#011validation-merror:0.442069\u001b[0m\n",
      "\u001b[34m[358]#011train-merror:0.262424#011validation-merror:0.441813\u001b[0m\n",
      "\u001b[34m[359]#011train-merror:0.262109#011validation-merror:0.441557\u001b[0m\n",
      "\u001b[34m[360]#011train-merror:0.261983#011validation-merror:0.441685\u001b[0m\n",
      "\u001b[34m[361]#011train-merror:0.261604#011validation-merror:0.441429\u001b[0m\n",
      "\u001b[34m[362]#011train-merror:0.260848#011validation-merror:0.441301\u001b[0m\n",
      "\u001b[34m[363]#011train-merror:0.260595#011validation-merror:0.441173\u001b[0m\n",
      "\u001b[34m[364]#011train-merror:0.260343#011validation-merror:0.441045\u001b[0m\n",
      "\u001b[34m[365]#011train-merror:0.260028#011validation-merror:0.441301\u001b[0m\n",
      "\u001b[34m[366]#011train-merror:0.260028#011validation-merror:0.441813\u001b[0m\n",
      "\u001b[34m[367]#011train-merror:0.259902#011validation-merror:0.441813\u001b[0m\n",
      "\u001b[34m[368]#011train-merror:0.259712#011validation-merror:0.442581\u001b[0m\n",
      "\u001b[34m[369]#011train-merror:0.259271#011validation-merror:0.442709\u001b[0m\n",
      "\u001b[34m[370]#011train-merror:0.258956#011validation-merror:0.442197\u001b[0m\n",
      "\u001b[34m[371]#011train-merror:0.258388#011validation-merror:0.442069\u001b[0m\n",
      "\u001b[34m[372]#011train-merror:0.258199#011validation-merror:0.442325\u001b[0m\n",
      "\u001b[34m[373]#011train-merror:0.258073#011validation-merror:0.441685\u001b[0m\n",
      "\u001b[34m[374]#011train-merror:0.258073#011validation-merror:0.441173\u001b[0m\n",
      "\u001b[34m[375]#011train-merror:0.257757#011validation-merror:0.441685\u001b[0m\n",
      "\u001b[34m[376]#011train-merror:0.257505#011validation-merror:0.441557\u001b[0m\n",
      "\u001b[34m[377]#011train-merror:0.257442#011validation-merror:0.442069\u001b[0m\n",
      "\u001b[34m[378]#011train-merror:0.256937#011validation-merror:0.441941\u001b[0m\n",
      "\u001b[34m[379]#011train-merror:0.256622#011validation-merror:0.442069\u001b[0m\n",
      "\u001b[34m[380]#011train-merror:0.256559#011validation-merror:0.442197\u001b[0m\n",
      "\u001b[34m[381]#011train-merror:0.256496#011validation-merror:0.442325\u001b[0m\n",
      "\u001b[34m[382]#011train-merror:0.256244#011validation-merror:0.442581\u001b[0m\n",
      "\u001b[34m[383]#011train-merror:0.256054#011validation-merror:0.442325\u001b[0m\n",
      "\u001b[34m[384]#011train-merror:0.255802#011validation-merror:0.442581\u001b[0m\n",
      "\u001b[34m[385]#011train-merror:0.255235#011validation-merror:0.443349\u001b[0m\n",
      "\u001b[34m[386]#011train-merror:0.255424#011validation-merror:0.442709\u001b[0m\n",
      "\u001b[34m[387]#011train-merror:0.254919#011validation-merror:0.443221\u001b[0m\n",
      "\u001b[34m[388]#011train-merror:0.254856#011validation-merror:0.443093\u001b[0m\n",
      "\u001b[34m[389]#011train-merror:0.254478#011validation-merror:0.443733\u001b[0m\n",
      "\u001b[34m[390]#011train-merror:0.254226#011validation-merror:0.443477\u001b[0m\n",
      "\u001b[34m[391]#011train-merror:0.254036#011validation-merror:0.443733\u001b[0m\n",
      "\u001b[34m[392]#011train-merror:0.253784#011validation-merror:0.443861\u001b[0m\n",
      "\u001b[34m[393]#011train-merror:0.253973#011validation-merror:0.444245\u001b[0m\n",
      "\u001b[34m[394]#011train-merror:0.253658#011validation-merror:0.444245\u001b[0m\n",
      "\u001b[34m[395]#011train-merror:0.253406#011validation-merror:0.444117\u001b[0m\n",
      "\u001b[34m[396]#011train-merror:0.253721#011validation-merror:0.444373\u001b[0m\n",
      "\u001b[34m[397]#011train-merror:0.253153#011validation-merror:0.444245\u001b[0m\n",
      "\u001b[34m[398]#011train-merror:0.252838#011validation-merror:0.444117\u001b[0m\n",
      "\u001b[34m[399]#011train-merror:0.252397#011validation-merror:0.444757\u001b[0m\n",
      "\u001b[34m[400]#011train-merror:0.252397#011validation-merror:0.444885\u001b[0m\n",
      "\u001b[34m[401]#011train-merror:0.25227#011validation-merror:0.445013\u001b[0m\n",
      "\u001b[34m[402]#011train-merror:0.252144#011validation-merror:0.445013\u001b[0m\n",
      "\u001b[34m[403]#011train-merror:0.252333#011validation-merror:0.444629\u001b[0m\n",
      "\u001b[34m[404]#011train-merror:0.251766#011validation-merror:0.444885\u001b[0m\n",
      "\u001b[34m[405]#011train-merror:0.251577#011validation-merror:0.444629\u001b[0m\n",
      "\u001b[34m[406]#011train-merror:0.251451#011validation-merror:0.444373\u001b[0m\n",
      "\u001b[34m[407]#011train-merror:0.251198#011validation-merror:0.444885\u001b[0m\n",
      "\u001b[34m[408]#011train-merror:0.251261#011validation-merror:0.445269\u001b[0m\n",
      "\u001b[34m[409]#011train-merror:0.251261#011validation-merror:0.444885\u001b[0m\n",
      "\u001b[34m[410]#011train-merror:0.251135#011validation-merror:0.445141\u001b[0m\n",
      "\u001b[34m[411]#011train-merror:0.250883#011validation-merror:0.445141\u001b[0m\n",
      "\u001b[34m[412]#011train-merror:0.250883#011validation-merror:0.444885\u001b[0m\n",
      "\u001b[34m[413]#011train-merror:0.250883#011validation-merror:0.445141\u001b[0m\n",
      "\u001b[34m[414]#011train-merror:0.25082#011validation-merror:0.444757\u001b[0m\n",
      "\u001b[34m[415]#011train-merror:0.250694#011validation-merror:0.444757\u001b[0m\n",
      "\u001b[34m[416]#011train-merror:0.250757#011validation-merror:0.444629\u001b[0m\n",
      "\u001b[34m[417]#011train-merror:0.250378#011validation-merror:0.444757\u001b[0m\n",
      "\u001b[34m[418]#011train-merror:0.250189#011validation-merror:0.444885\u001b[0m\n",
      "\u001b[34m[419]#011train-merror:0.250063#011validation-merror:0.445526\u001b[0m\n",
      "\u001b[34m[420]#011train-merror:0.249937#011validation-merror:0.445141\u001b[0m\n",
      "\u001b[34m[421]#011train-merror:0.249811#011validation-merror:0.445269\u001b[0m\n",
      "\u001b[34m[422]#011train-merror:0.249748#011validation-merror:0.444757\u001b[0m\n",
      "\u001b[34m[423]#011train-merror:0.249432#011validation-merror:0.445013\u001b[0m\n",
      "\u001b[34m[424]#011train-merror:0.249622#011validation-merror:0.444117\u001b[0m\n",
      "\u001b[34m[425]#011train-merror:0.249432#011validation-merror:0.444757\u001b[0m\n",
      "\u001b[34m[426]#011train-merror:0.249117#011validation-merror:0.445526\u001b[0m\n",
      "\u001b[34m[427]#011train-merror:0.249495#011validation-merror:0.445269\u001b[0m\n",
      "\u001b[34m[428]#011train-merror:0.248549#011validation-merror:0.445398\u001b[0m\n",
      "\u001b[34m[429]#011train-merror:0.248045#011validation-merror:0.445141\u001b[0m\n",
      "\u001b[34m[430]#011train-merror:0.247793#011validation-merror:0.444501\u001b[0m\n",
      "\u001b[34m[431]#011train-merror:0.247666#011validation-merror:0.444885\u001b[0m\n",
      "\u001b[34m[432]#011train-merror:0.247666#011validation-merror:0.444629\u001b[0m\n",
      "\u001b[34m[433]#011train-merror:0.247414#011validation-merror:0.444757\u001b[0m\n",
      "\u001b[34m[434]#011train-merror:0.247414#011validation-merror:0.444757\u001b[0m\n",
      "\u001b[34m[435]#011train-merror:0.247414#011validation-merror:0.445269\u001b[0m\n",
      "\u001b[34m[436]#011train-merror:0.247099#011validation-merror:0.445269\u001b[0m\n",
      "\u001b[34m[437]#011train-merror:0.247225#011validation-merror:0.445526\u001b[0m\n",
      "\u001b[34m[438]#011train-merror:0.246973#011validation-merror:0.445398\u001b[0m\n",
      "\u001b[34m[439]#011train-merror:0.24691#011validation-merror:0.445654\u001b[0m\n",
      "\u001b[34m[440]#011train-merror:0.24672#011validation-merror:0.445398\u001b[0m\n",
      "\u001b[34m[441]#011train-merror:0.246657#011validation-merror:0.445526\u001b[0m\n",
      "\u001b[34m[442]#011train-merror:0.246342#011validation-merror:0.445013\u001b[0m\n",
      "\u001b[34m[443]#011train-merror:0.246153#011validation-merror:0.444757\u001b[0m\n",
      "\u001b[34m[444]#011train-merror:0.246027#011validation-merror:0.444757\u001b[0m\n",
      "\u001b[34m[445]#011train-merror:0.246027#011validation-merror:0.444117\u001b[0m\n",
      "\u001b[34m[446]#011train-merror:0.245964#011validation-merror:0.443477\u001b[0m\n",
      "\u001b[34m[447]#011train-merror:0.245459#011validation-merror:0.443733\u001b[0m\n",
      "\u001b[34m[448]#011train-merror:0.245585#011validation-merror:0.443605\u001b[0m\n",
      "\u001b[34m[449]#011train-merror:0.245838#011validation-merror:0.443605\u001b[0m\n",
      "\u001b[34m[450]#011train-merror:0.245522#011validation-merror:0.443477\u001b[0m\n",
      "\u001b[34m[451]#011train-merror:0.245081#011validation-merror:0.443093\u001b[0m\n",
      "\u001b[34m[452]#011train-merror:0.244702#011validation-merror:0.442965\u001b[0m\n",
      "\u001b[34m[453]#011train-merror:0.244387#011validation-merror:0.443221\u001b[0m\n",
      "\u001b[34m[454]#011train-merror:0.243819#011validation-merror:0.442837\u001b[0m\n",
      "\u001b[34m[455]#011train-merror:0.243567#011validation-merror:0.443221\u001b[0m\n",
      "\u001b[34m[456]#011train-merror:0.243567#011validation-merror:0.443349\u001b[0m\n",
      "\u001b[34m[457]#011train-merror:0.243189#011validation-merror:0.443349\u001b[0m\n",
      "\u001b[34m[458]#011train-merror:0.243063#011validation-merror:0.444117\u001b[0m\n",
      "\u001b[34m[459]#011train-merror:0.243126#011validation-merror:0.444117\u001b[0m\n",
      "\u001b[34m[460]#011train-merror:0.243189#011validation-merror:0.443861\u001b[0m\n",
      "\u001b[34m[461]#011train-merror:0.242936#011validation-merror:0.444629\u001b[0m\n",
      "\u001b[34m[462]#011train-merror:0.242747#011validation-merror:0.445013\u001b[0m\n",
      "\u001b[34m[463]#011train-merror:0.242747#011validation-merror:0.445398\u001b[0m\n",
      "\u001b[34m[464]#011train-merror:0.242999#011validation-merror:0.445526\u001b[0m\n",
      "\u001b[34m[465]#011train-merror:0.242558#011validation-merror:0.445398\u001b[0m\n",
      "\u001b[34m[466]#011train-merror:0.242684#011validation-merror:0.445141\u001b[0m\n",
      "\u001b[34m[467]#011train-merror:0.242306#011validation-merror:0.445782\u001b[0m\n",
      "\u001b[34m[468]#011train-merror:0.24199#011validation-merror:0.445526\u001b[0m\n",
      "\u001b[34m[469]#011train-merror:0.241927#011validation-merror:0.445398\u001b[0m\n",
      "\u001b[34m[470]#011train-merror:0.241864#011validation-merror:0.445526\u001b[0m\n",
      "\u001b[34m[471]#011train-merror:0.241612#011validation-merror:0.445526\u001b[0m\n",
      "\u001b[34m[472]#011train-merror:0.240729#011validation-merror:0.446294\u001b[0m\n",
      "\u001b[34m[473]#011train-merror:0.240666#011validation-merror:0.446678\u001b[0m\n",
      "\u001b[34m[474]#011train-merror:0.240666#011validation-merror:0.446806\u001b[0m\n",
      "\u001b[34m[475]#011train-merror:0.240098#011validation-merror:0.446934\u001b[0m\n",
      "\u001b[34m[476]#011train-merror:0.240098#011validation-merror:0.447062\u001b[0m\n",
      "\u001b[34m[477]#011train-merror:0.239657#011validation-merror:0.44655\u001b[0m\n",
      "\u001b[34m[478]#011train-merror:0.239531#011validation-merror:0.447318\u001b[0m\n",
      "\u001b[34m[479]#011train-merror:0.239215#011validation-merror:0.446934\u001b[0m\n",
      "\u001b[34m[480]#011train-merror:0.238837#011validation-merror:0.447062\u001b[0m\n",
      "\u001b[34m[481]#011train-merror:0.238648#011validation-merror:0.44719\u001b[0m\n",
      "\u001b[34m[482]#011train-merror:0.238206#011validation-merror:0.446422\u001b[0m\n",
      "\u001b[34m[483]#011train-merror:0.238396#011validation-merror:0.446806\u001b[0m\n",
      "\u001b[34m[484]#011train-merror:0.237954#011validation-merror:0.447574\u001b[0m\n",
      "\u001b[34m[485]#011train-merror:0.237513#011validation-merror:0.447318\u001b[0m\n",
      "\u001b[34m[486]#011train-merror:0.237071#011validation-merror:0.447446\u001b[0m\n",
      "\u001b[34m[487]#011train-merror:0.23663#011validation-merror:0.447958\u001b[0m\n",
      "\u001b[34m[488]#011train-merror:0.236819#011validation-merror:0.447446\u001b[0m\n",
      "\u001b[34m[489]#011train-merror:0.23663#011validation-merror:0.44783\u001b[0m\n",
      "\u001b[34m[490]#011train-merror:0.236567#011validation-merror:0.448214\u001b[0m\n",
      "\u001b[34m[491]#011train-merror:0.236567#011validation-merror:0.448342\u001b[0m\n",
      "\u001b[34m[492]#011train-merror:0.236504#011validation-merror:0.448342\u001b[0m\n",
      "\u001b[34m[493]#011train-merror:0.236567#011validation-merror:0.448726\u001b[0m\n",
      "\u001b[34m[494]#011train-merror:0.236567#011validation-merror:0.44847\u001b[0m\n",
      "\u001b[34m[495]#011train-merror:0.23663#011validation-merror:0.448342\u001b[0m\n",
      "\u001b[34m[496]#011train-merror:0.236188#011validation-merror:0.448598\u001b[0m\n",
      "\u001b[34m[497]#011train-merror:0.236188#011validation-merror:0.448726\u001b[0m\n",
      "\u001b[34m[498]#011train-merror:0.236062#011validation-merror:0.44847\u001b[0m\n",
      "\u001b[34m[499]#011train-merror:0.235431#011validation-merror:0.44847\u001b[0m\n",
      "\n",
      "2020-02-01 16:58:08 Uploading - Uploading generated training model\n",
      "2020-02-01 16:58:08 Completed - Training job completed\n",
      "Training seconds: 72\n",
      "Billable seconds: 72\n"
     ]
    }
   ],
   "source": [
    "s3_input_train = sagemaker.s3_input(s3_data=train_location, content_type='csv')\n",
    "s3_input_validation = sagemaker.s3_input(s3_data=val_location, content_type='csv')\n",
    "\n",
    "xgb.fit({'train': s3_input_train, 'validation': s3_input_validation})"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Test default model performance"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Build a transformer object from our fit model:\n",
    "xgb_transformer = xgb.transformer(instance_count = 1, instance_type = 'ml.m4.xlarge')\n",
    "\n",
    "# Begin a batch transform job using our trained model and applying it to the test data:\n",
    "xgb_transformer.transform(test_location, content_type='text/csv', split_type='Line')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "......................\u001b[34m[2020-02-01 17:02:07 +0000] [14] [INFO] Starting gunicorn 19.10.0\u001b[0m\n",
      "\u001b[34m[2020-02-01 17:02:07 +0000] [14] [INFO] Listening at: unix:/tmp/gunicorn.sock (14)\u001b[0m\n",
      "\u001b[34m[2020-02-01 17:02:07 +0000] [14] [INFO] Using worker: gevent\u001b[0m\n",
      "\u001b[34m[2020-02-01 17:02:07 +0000] [21] [INFO] Booting worker with pid: 21\u001b[0m\n",
      "\u001b[34m[2020-02-01 17:02:07 +0000] [25] [INFO] Booting worker with pid: 25\u001b[0m\n",
      "\u001b[34m[2020-02-01 17:02:08 +0000] [26] [INFO] Booting worker with pid: 26\u001b[0m\n",
      "\u001b[34m[2020-02-01 17:02:08 +0000] [27] [INFO] Booting worker with pid: 27\u001b[0m\n",
      "\n",
      "\u001b[34m[2020-02-01:17:02:37:INFO] No GPUs detected (normal if no gpus installed)\u001b[0m\n",
      "\u001b[34m169.254.255.130 - - [01/Feb/2020:17:02:37 +0000] \"GET /ping HTTP/1.1\" 200 0 \"-\" \"Go-http-client/1.1\"\u001b[0m\n",
      "\u001b[34m[2020-02-01:17:02:37:INFO] No GPUs detected (normal if no gpus installed)\u001b[0m\n",
      "\u001b[34m169.254.255.130 - - [01/Feb/2020:17:02:37 +0000] \"GET /execution-parameters HTTP/1.1\" 200 84 \"-\" \"Go-http-client/1.1\"\u001b[0m\n",
      "\u001b[34m[2020-02-01:17:02:37:INFO] No GPUs detected (normal if no gpus installed)\u001b[0m\n",
      "\u001b[34m[2020-02-01:17:02:37:INFO] Determined delimiter of CSV input is ','\u001b[0m\n",
      "\u001b[34m169.254.255.130 - - [01/Feb/2020:17:02:39 +0000] \"POST /invocations HTTP/1.1\" 200 731721 \"-\" \"Go-http-client/1.1\"\u001b[0m\n",
      "\u001b[32m2020-02-01T17:02:37.336:[sagemaker logs]: MaxConcurrentTransforms=4, MaxPayloadInMB=6, BatchStrategy=MULTI_RECORD\u001b[0m\n"
     ]
    }
   ],
   "source": [
    "xgb_transformer.wait()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "download: s3://sagemaker-us-east-1-037802178319/sagemaker-xgboost-2020-02-01-16-58-41-381/test.csv.out to ../data/test.csv.out\r\n"
     ]
    }
   ],
   "source": [
    "!aws s3 cp --recursive $xgb_transformer.output_path $data_dir"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>0</th>\n",
       "      <th>1</th>\n",
       "      <th>2</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>[0.07351533323526382</td>\n",
       "      <td>0.026411</td>\n",
       "      <td>0.9000735282897949]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>[0.06258437782526016</td>\n",
       "      <td>0.595232</td>\n",
       "      <td>0.34218382835388184]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>[0.04993101581931114</td>\n",
       "      <td>0.057054</td>\n",
       "      <td>0.8930152058601379]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>[0.07440779358148575</td>\n",
       "      <td>0.323170</td>\n",
       "      <td>0.6024218797683716]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>[0.03284726291894913</td>\n",
       "      <td>0.075518</td>\n",
       "      <td>0.8916345834732056]</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                      0         1                      2\n",
       "0  [0.07351533323526382  0.026411    0.9000735282897949]\n",
       "1  [0.06258437782526016  0.595232   0.34218382835388184]\n",
       "2  [0.04993101581931114  0.057054    0.8930152058601379]\n",
       "3  [0.07440779358148575  0.323170    0.6024218797683716]\n",
       "4  [0.03284726291894913  0.075518    0.8916345834732056]"
      ]
     },
     "execution_count": 25,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y_pred = pd.read_csv(os.path.join(data_dir, 'test.csv.out'), header=None)\n",
    "y_pred.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>0</th>\n",
       "      <th>1</th>\n",
       "      <th>2</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0.073515</td>\n",
       "      <td>0.026411</td>\n",
       "      <td>0.900074</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0.062584</td>\n",
       "      <td>0.595232</td>\n",
       "      <td>0.342184</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0.049931</td>\n",
       "      <td>0.057054</td>\n",
       "      <td>0.893015</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0.074408</td>\n",
       "      <td>0.323170</td>\n",
       "      <td>0.602422</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0.032847</td>\n",
       "      <td>0.075518</td>\n",
       "      <td>0.891635</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "          0         1         2\n",
       "0  0.073515  0.026411  0.900074\n",
       "1  0.062584  0.595232  0.342184\n",
       "2  0.049931  0.057054  0.893015\n",
       "3  0.074408  0.323170  0.602422\n",
       "4  0.032847  0.075518  0.891635"
      ]
     },
     "execution_count": 26,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Fix the brackets and convert string to a float:\n",
    "y_pred[0] = y_pred[0].str.strip('[').astype(float)\n",
    "y_pred[2] = y_pred[2].str.strip(']').astype(float)\n",
    "\n",
    "y_pred.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>0</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   0\n",
       "0  2\n",
       "1  1\n",
       "2  2\n",
       "3  2\n",
       "4  2"
      ]
     },
     "execution_count": 27,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Choose the most likely output based on predicted probabilities:\n",
    "y_max = pd.DataFrame()\n",
    "y_max[0] = y_pred[[0, 1, 2]].idxmax(axis=1)\n",
    "\n",
    "y_max.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Default model accuracy score: 0.5519\n"
     ]
    }
   ],
   "source": [
    "from sklearn.metrics import accuracy_score\n",
    "\n",
    "print(\"Default model accuracy score: {:.4f}\".format(accuracy_score(y_test, y_max)))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Model refinement\n",
    "#### Create the hyperparameter tuner\n",
    "\n",
    "Amazon SageMaker automatic model tuning, also known as hyperparameter tuning, finds the best version of a model by running many training jobs on your dataset using the algorithm and ranges of hyperparameters that you specify. It then chooses the hyperparameter values that result in a model that performs the best, as measured by a metric that you choose (from https://docs.aws.amazon.com/sagemaker/latest/dg/automatic-model-tuning.html).\n",
    "\n",
    "To compare the trained models, we will use **merror**: multiclass classification error rate. It is calculated as #(wrong cases)/#(all cases).\n",
    "\n",
    "We will use following hyperparameters to refine the initial model:\n",
    "(from https://xgboost.readthedocs.io/en/latest/parameter.html)\n",
    "\n",
    "- **eta**: Step size shrinkage used in update to prevents overfitting. After each boosting step, we can directly get the weights of new features, and eta shrinks the feature weights to make the boosting process more conservative.\n",
    "\n",
    "- **gamma**: Minimum loss reduction required to make a further partition on a leaf node of the tree. The larger gamma is, the more conservative the algorithm will be.\n",
    "\n",
    "- **max_depth**: Maximum depth of a tree. Increasing this value will make the model more complex and more likely to overfit.\n",
    "\n",
    "- **subsample**: Subsample ratio of the training instances. Setting it to 0.5 means that XGBoost would randomly sample half of the training data prior to growing trees. and this will prevent overfitting. Subsampling will occur once in every boosting iteration"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sagemaker.tuner import IntegerParameter, ContinuousParameter, HyperparameterTuner\n",
    "\n",
    "xgb_hyperparameter_tuner = HyperparameterTuner(estimator = xgb, # the estimator object to use as the basis\n",
    "                                               objective_metric_name = 'validation:merror', # the metric used to compare trained models\n",
    "                                               objective_type = 'Minimize', # whether we wish to minimize or maximize the metric\n",
    "                                               max_jobs = 32, # the total number of models to train\n",
    "                                               max_parallel_jobs = 3, # the number of models to train in parallel\n",
    "                                               hyperparameter_ranges = {\n",
    "                                                    'eta'      : ContinuousParameter(0.05, 0.5),\n",
    "                                                    'gamma': ContinuousParameter(0, 10),\n",
    "                                                    'max_depth': IntegerParameter(2, 8),\n",
    "                                                    'subsample': ContinuousParameter(0.5, 1),\n",
    "                                               })"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Find the best version of a model by running many training jobs"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [],
   "source": [
    "xgb_hyperparameter_tuner.fit({'train': s3_input_train, 'validation': s3_input_validation})"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      ".............................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................!\n"
     ]
    }
   ],
   "source": [
    "xgb_hyperparameter_tuner.wait()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Retrieve and set up the best performing model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "'sagemaker-xgboost-200201-1703-001-0e60ab7e'"
      ]
     },
     "execution_count": 32,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "xgb_hyperparameter_tuner.best_training_job()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "2020-02-01 17:06:30 Starting - Preparing the instances for training\n",
      "2020-02-01 17:06:30 Downloading - Downloading input data\n",
      "2020-02-01 17:06:30 Training - Training image download completed. Training in progress.\n",
      "2020-02-01 17:06:30 Uploading - Uploading generated training model\n",
      "2020-02-01 17:06:30 Completed - Training job completed\u001b[34mINFO:sagemaker-containers:Imported framework sagemaker_xgboost_container.training\u001b[0m\n",
      "\u001b[34mINFO:sagemaker-containers:Failed to parse hyperparameter _tuning_objective_metric value validation:merror to Json.\u001b[0m\n",
      "\u001b[34mReturning the value itself\u001b[0m\n",
      "\u001b[34mINFO:sagemaker-containers:Failed to parse hyperparameter objective value multi:softprob to Json.\u001b[0m\n",
      "\u001b[34mReturning the value itself\u001b[0m\n",
      "\u001b[34mINFO:sagemaker-containers:No GPUs detected (normal if no gpus installed)\u001b[0m\n",
      "\u001b[34mINFO:sagemaker_xgboost_container.training:Running XGBoost Sagemaker in algorithm mode\u001b[0m\n",
      "\u001b[34mINFO:root:Determined delimiter of CSV input is ','\u001b[0m\n",
      "\u001b[34mINFO:root:Determined delimiter of CSV input is ','\u001b[0m\n",
      "\u001b[34mINFO:root:Determined delimiter of CSV input is ','\u001b[0m\n",
      "\u001b[34m[17:06:13] 15856x8 matrix with 126848 entries loaded from /opt/ml/input/data/train?format=csv&label_column=0&delimiter=,\u001b[0m\n",
      "\u001b[34mINFO:root:Determined delimiter of CSV input is ','\u001b[0m\n",
      "\u001b[34m[17:06:13] 7811x8 matrix with 62488 entries loaded from /opt/ml/input/data/validation?format=csv&label_column=0&delimiter=,\u001b[0m\n",
      "\u001b[34mINFO:root:Single node training.\u001b[0m\n",
      "\u001b[34mINFO:root:Setting up HPO optimized metric to be : merror\u001b[0m\n",
      "\u001b[34mINFO:root:Train matrix has 15856 rows\u001b[0m\n",
      "\u001b[34mINFO:root:Validation matrix has 7811 rows\u001b[0m\n",
      "\u001b[34m[0]#011train-merror:0.395119#011validation-merror:0.395084\u001b[0m\n",
      "\u001b[34m[1]#011train-merror:0.395119#011validation-merror:0.395084\u001b[0m\n",
      "\u001b[34m[2]#011train-merror:0.395119#011validation-merror:0.394956\u001b[0m\n",
      "\u001b[34m[3]#011train-merror:0.395119#011validation-merror:0.394956\u001b[0m\n",
      "\u001b[34m[4]#011train-merror:0.395119#011validation-merror:0.395084\u001b[0m\n",
      "\u001b[34m[5]#011train-merror:0.395119#011validation-merror:0.394956\u001b[0m\n",
      "\u001b[34m[6]#011train-merror:0.395119#011validation-merror:0.394956\u001b[0m\n",
      "\u001b[34m[7]#011train-merror:0.395812#011validation-merror:0.39534\u001b[0m\n",
      "\u001b[34m[8]#011train-merror:0.395119#011validation-merror:0.395084\u001b[0m\n",
      "\u001b[34m[9]#011train-merror:0.395119#011validation-merror:0.395084\u001b[0m\n",
      "\u001b[34m[10]#011train-merror:0.395119#011validation-merror:0.394956\u001b[0m\n",
      "\u001b[34m[11]#011train-merror:0.395119#011validation-merror:0.395084\u001b[0m\n",
      "\u001b[34m[12]#011train-merror:0.395119#011validation-merror:0.395084\u001b[0m\n",
      "\u001b[34m[13]#011train-merror:0.395119#011validation-merror:0.395084\u001b[0m\n",
      "\u001b[34m[14]#011train-merror:0.395119#011validation-merror:0.395084\u001b[0m\n",
      "\u001b[34m[15]#011train-merror:0.395119#011validation-merror:0.395084\u001b[0m\n",
      "\u001b[34m[16]#011train-merror:0.395119#011validation-merror:0.395084\u001b[0m\n",
      "\u001b[34m[17]#011train-merror:0.395119#011validation-merror:0.395084\u001b[0m\n",
      "\u001b[34m[18]#011train-merror:0.395119#011validation-merror:0.395084\u001b[0m\n",
      "\u001b[34m[19]#011train-merror:0.395182#011validation-merror:0.394316\u001b[0m\n",
      "\u001b[34m[20]#011train-merror:0.395182#011validation-merror:0.394316\u001b[0m\n",
      "\u001b[34m[21]#011train-merror:0.395182#011validation-merror:0.394316\u001b[0m\n",
      "\u001b[34m[22]#011train-merror:0.395182#011validation-merror:0.394316\u001b[0m\n",
      "\u001b[34m[23]#011train-merror:0.395182#011validation-merror:0.394316\u001b[0m\n",
      "\u001b[34m[24]#011train-merror:0.395182#011validation-merror:0.394316\u001b[0m\n",
      "\u001b[34m[25]#011train-merror:0.395182#011validation-merror:0.394316\u001b[0m\n",
      "\u001b[34m[26]#011train-merror:0.395182#011validation-merror:0.394316\u001b[0m\n",
      "\u001b[34m[27]#011train-merror:0.395182#011validation-merror:0.394316\u001b[0m\n",
      "\u001b[34m[28]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[29]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[30]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[31]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[32]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[33]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[34]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[35]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[36]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[37]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[38]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[39]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[40]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[41]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[42]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[43]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[44]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[45]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[46]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[47]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[48]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[49]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[50]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[51]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[52]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[53]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[54]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[55]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[56]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[57]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[58]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[59]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[60]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[61]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[62]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[63]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[64]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[65]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[66]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[67]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[68]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[69]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[70]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[71]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[72]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[73]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[74]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[75]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[76]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[77]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[78]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[79]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[80]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[81]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[82]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[83]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[84]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[85]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[86]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[87]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[88]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[89]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[90]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[91]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[92]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[93]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[94]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[95]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[96]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[97]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[98]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[99]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[100]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[101]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[102]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[103]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[104]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[105]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[106]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[107]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[108]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[109]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[110]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[111]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[112]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[113]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[114]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[115]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[116]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[117]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[118]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[119]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[120]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[121]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[122]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[123]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[124]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[125]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[126]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[127]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[128]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[129]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[130]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[131]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[132]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[133]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[134]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[135]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[136]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[137]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[138]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[139]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[140]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[141]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[142]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[143]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[144]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[145]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[146]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[147]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[148]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[149]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[150]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[151]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[152]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[153]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[154]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[155]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[156]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[157]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[158]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[159]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[160]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[161]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[162]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[163]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[164]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[165]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[166]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[167]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[168]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[169]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[170]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[171]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[172]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[173]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[174]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[175]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[176]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[177]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[178]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[179]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[180]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[181]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[182]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[183]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[184]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[185]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[186]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[187]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[188]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[189]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[190]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[191]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[192]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[193]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[194]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[195]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[196]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[197]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[198]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[199]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[200]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[201]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[202]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[203]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[204]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[205]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[206]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[207]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[208]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[209]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[210]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[211]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[212]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[213]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[214]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[215]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[216]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[217]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[218]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[219]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[220]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[221]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[222]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[223]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[224]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[225]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[226]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[227]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[228]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[229]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[230]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[231]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[232]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[233]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[234]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[235]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[236]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[237]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[238]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[239]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[240]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[241]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[242]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[243]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[244]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[245]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[246]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[247]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[248]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[249]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[250]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[251]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[252]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[253]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[254]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[255]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[256]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[257]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[258]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[259]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[260]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[261]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[262]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[263]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[264]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[265]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[266]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[267]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[268]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[269]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[270]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[271]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[272]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[273]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[274]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[275]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[276]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[277]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[278]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[279]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[280]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[281]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[282]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[283]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[284]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[285]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[286]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[287]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[288]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[289]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[290]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[291]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[292]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[293]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[294]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[295]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[296]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[297]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[298]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[299]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[300]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[301]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[302]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[303]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[304]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[305]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[306]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[307]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[308]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[309]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[310]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[311]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[312]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[313]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[314]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[315]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[316]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[317]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[318]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[319]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[320]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[321]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[322]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[323]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[324]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[325]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[326]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[327]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[328]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[329]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[330]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[331]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[332]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[333]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[334]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[335]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[336]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[337]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[338]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[339]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[340]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[341]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[342]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[343]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[344]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[345]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[346]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[347]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[348]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[349]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[350]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[351]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[352]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[353]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[354]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[355]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[356]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[357]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[358]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[359]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[360]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[361]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[362]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[363]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[364]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[365]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[366]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[367]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[368]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[369]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[370]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[371]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[372]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[373]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[374]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[375]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[376]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[377]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[378]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[379]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[380]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[381]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[382]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[383]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[384]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[385]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[386]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[387]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[388]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[389]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[390]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[391]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[392]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[393]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[394]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[395]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[396]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[397]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[398]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[399]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[400]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[401]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[402]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[403]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[404]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[405]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[406]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[407]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[408]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[409]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[410]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[411]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[412]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[413]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[414]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[415]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[416]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[417]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[418]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[419]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[420]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[421]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[422]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[423]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[424]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[425]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[426]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[427]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[428]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[429]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[430]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[431]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[432]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[433]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[434]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[435]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[436]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[437]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[438]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[439]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[440]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[441]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[442]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[443]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[444]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[445]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[446]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[447]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[448]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[449]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[450]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[451]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[452]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[453]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[454]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[455]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[456]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[457]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[458]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[459]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[460]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[461]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[462]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[463]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[464]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[465]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[466]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[467]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[468]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[469]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[470]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[471]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[472]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[473]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[474]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[475]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[476]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[477]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[478]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[479]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[480]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[481]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[482]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[483]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[484]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[485]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[486]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[487]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[488]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[489]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[490]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[491]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[492]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[493]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[494]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[495]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[496]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[497]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[498]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "\u001b[34m[499]#011train-merror:0.395245#011validation-merror:0.394188\u001b[0m\n",
      "Training seconds: 68\n",
      "Billable seconds: 68\n"
     ]
    }
   ],
   "source": [
    "# Set up a batch transform job to test the best model:\n",
    "xgb_attached = sagemaker.estimator.Estimator.attach(xgb_hyperparameter_tuner.best_training_job())"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Test new model performance"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Build a transformer object from our fit model:\n",
    "xgb_transformer = xgb_attached.transformer(instance_count = 1, instance_type = 'ml.m4.xlarge')\n",
    "\n",
    "# Begin a batch transform job using our trained model and applying it to the test data:\n",
    "xgb_transformer.transform(test_location, content_type='text/csv', split_type='Line')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      ".......................\u001b[34m[2020-02-01 17:51:20 +0000] [14] [INFO] Starting gunicorn 19.10.0\u001b[0m\n",
      "\u001b[34m[2020-02-01 17:51:20 +0000] [14] [INFO] Listening at: unix:/tmp/gunicorn.sock (14)\u001b[0m\n",
      "\u001b[34m[2020-02-01 17:51:20 +0000] [14] [INFO] Using worker: gevent\u001b[0m\n",
      "\u001b[34m[2020-02-01 17:51:20 +0000] [21] [INFO] Booting worker with pid: 21\u001b[0m\n",
      "\u001b[34m[2020-02-01 17:51:20 +0000] [22] [INFO] Booting worker with pid: 22\u001b[0m\n",
      "\u001b[34m[2020-02-01 17:51:20 +0000] [23] [INFO] Booting worker with pid: 23\u001b[0m\n",
      "\u001b[34m[2020-02-01 17:51:20 +0000] [30] [INFO] Booting worker with pid: 30\u001b[0m\n",
      "\n",
      "\u001b[34m[2020-02-01:17:51:42:INFO] No GPUs detected (normal if no gpus installed)\u001b[0m\n",
      "\u001b[34m169.254.255.130 - - [01/Feb/2020:17:51:42 +0000] \"GET /ping HTTP/1.1\" 200 0 \"-\" \"Go-http-client/1.1\"\u001b[0m\n",
      "\u001b[34m[2020-02-01:17:51:42:INFO] No GPUs detected (normal if no gpus installed)\u001b[0m\n",
      "\u001b[34m169.254.255.130 - - [01/Feb/2020:17:51:42 +0000] \"GET /execution-parameters HTTP/1.1\" 200 84 \"-\" \"Go-http-client/1.1\"\u001b[0m\n",
      "\u001b[34m[2020-02-01:17:51:43:INFO] Determined delimiter of CSV input is ','\u001b[0m\n",
      "\u001b[35m[2020-02-01:17:51:42:INFO] No GPUs detected (normal if no gpus installed)\u001b[0m\n",
      "\u001b[35m169.254.255.130 - - [01/Feb/2020:17:51:42 +0000] \"GET /ping HTTP/1.1\" 200 0 \"-\" \"Go-http-client/1.1\"\u001b[0m\n",
      "\u001b[35m[2020-02-01:17:51:42:INFO] No GPUs detected (normal if no gpus installed)\u001b[0m\n",
      "\u001b[35m169.254.255.130 - - [01/Feb/2020:17:51:42 +0000] \"GET /execution-parameters HTTP/1.1\" 200 84 \"-\" \"Go-http-client/1.1\"\u001b[0m\n",
      "\u001b[35m[2020-02-01:17:51:43:INFO] Determined delimiter of CSV input is ','\u001b[0m\n",
      "\u001b[34m169.254.255.130 - - [01/Feb/2020:17:51:43 +0000] \"POST /invocations HTTP/1.1\" 200 724164 \"-\" \"Go-http-client/1.1\"\u001b[0m\n",
      "\u001b[35m169.254.255.130 - - [01/Feb/2020:17:51:43 +0000] \"POST /invocations HTTP/1.1\" 200 724164 \"-\" \"Go-http-client/1.1\"\u001b[0m\n",
      "\u001b[32m2020-02-01T17:51:42.963:[sagemaker logs]: MaxConcurrentTransforms=4, MaxPayloadInMB=6, BatchStrategy=MULTI_RECORD\u001b[0m\n"
     ]
    }
   ],
   "source": [
    "xgb_transformer.wait()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "download: s3://sagemaker-us-east-1-037802178319/sagemaker-xgboost-200201-1703-001-0e60a-2020-02-01-17-47-49-061/test.csv.out to ../data/test.csv.out\r\n"
     ]
    }
   ],
   "source": [
    "!aws s3 cp --recursive $xgb_transformer.output_path $data_dir"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Tuned model accuracy score: 0.6050\n"
     ]
    }
   ],
   "source": [
    "y_pred = pd.read_csv(os.path.join(data_dir, 'test.csv.out'), header=None)\n",
    "y_pred[0] = y_pred[0].str.strip('[').astype(float)\n",
    "y_pred[2] = y_pred[2].str.strip(']').astype(float)\n",
    "\n",
    "y_max = pd.DataFrame()\n",
    "y_max[0] = y_pred[[0, 1, 2]].idxmax(axis=1)\n",
    "\n",
    "print(\"Tuned model accuracy score: {:.4f}\".format(accuracy_score(y_test, y_max)))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "                precision    recall  f1-score   support\n",
      "\n",
      "       no view       0.00      0.00      0.00      1606\n",
      "      no order       0.58      0.18      0.28      3224\n",
      "possible order       0.61      0.95      0.74      6827\n",
      "\n",
      "     micro avg       0.60      0.60      0.60     11657\n",
      "     macro avg       0.39      0.38      0.34     11657\n",
      "  weighted avg       0.52      0.60      0.51     11657\n",
      "\n"
     ]
    }
   ],
   "source": [
    "from sklearn.metrics import classification_report\n",
    "\n",
    "labels = [\"no view\", \"no order\", \"possible order\"]\n",
    "print(classification_report(y_test, y_max, target_names=labels))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 576x576 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "from sklearn.metrics import confusion_matrix\n",
    "\n",
    "cm = confusion_matrix(y_test, y_max)\n",
    "\n",
    "f, ax = plt.subplots(figsize = (8,8))\n",
    "\n",
    "sns.heatmap(cm,\n",
    "            cmap=sns.cubehelix_palette(8, start=.5, rot=-.75),\n",
    "            square=True,\n",
    "            annot=True, annot_kws={\"size\": 12}, fmt='g',\n",
    "            xticklabels=labels,\n",
    "            yticklabels=labels)\n",
    "\n",
    "plt.title('Confusion matrix of the XGBoost classifier\\n\\n\\n', fontsize=16)\n",
    "\n",
    "plt.xlabel(\"\\npredicted label\")\n",
    "plt.ylabel(\"true label\\n\")\n",
    "\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Deploy the XGBoost model\n",
    "#### Create a simulated data set with new customers\n",
    "The purpose of this model is to be able to use existing data to target **new customers** with offers. When new customer registers in the Starbucks Rewards® program, we can feed it's data to the deployed model and draw a conclusions about possible offer types for this customer.\n",
    "\n",
    "We don't have real data about new customers, so we will generate this data."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>age</th>\n",
       "      <th>income</th>\n",
       "      <th>gender_F</th>\n",
       "      <th>gender_M</th>\n",
       "      <th>gender_O</th>\n",
       "      <th>bogo</th>\n",
       "      <th>discount</th>\n",
       "      <th>informational</th>\n",
       "      <th>person</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>20</td>\n",
       "      <td>75724</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>24</td>\n",
       "      <td>118385</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>47</td>\n",
       "      <td>29441</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>59</td>\n",
       "      <td>33835</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>18</td>\n",
       "      <td>119267</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>37</td>\n",
       "      <td>85146</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>32</td>\n",
       "      <td>70816</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>6</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>66</td>\n",
       "      <td>47565</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>7</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>76</td>\n",
       "      <td>115821</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>8</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>24</td>\n",
       "      <td>54680</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>9</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   age  income  gender_F  gender_M  gender_O  bogo  discount  informational  \\\n",
       "0   20   75724         0         0         1     0         0              0   \n",
       "1   24  118385         0         0         1     0         0              0   \n",
       "2   47   29441         1         0         0     0         0              0   \n",
       "3   59   33835         0         1         0     0         0              0   \n",
       "4   18  119267         1         0         0     0         0              0   \n",
       "5   37   85146         1         0         0     0         0              0   \n",
       "6   32   70816         0         0         1     0         0              0   \n",
       "7   66   47565         0         1         0     0         0              0   \n",
       "8   76  115821         0         1         0     0         0              0   \n",
       "9   24   54680         0         0         1     0         0              0   \n",
       "\n",
       "   person  \n",
       "0       0  \n",
       "1       1  \n",
       "2       2  \n",
       "3       3  \n",
       "4       4  \n",
       "5       5  \n",
       "6       6  \n",
       "7       7  \n",
       "8       8  \n",
       "9       9  "
      ]
     },
     "execution_count": 40,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "new_customers = pd.DataFrame()\n",
    "\n",
    "new_customers.colums = ['age', 'income', 'gender', 'bogo', 'discount', 'informational']\n",
    "new_customers['age'] = np.random.randint(18, 85, 10) # random ages from 18 to 85\n",
    "new_customers['income'] = np.random.randint(20000, 120000, 10) # random income from 20000 to 120000\n",
    "\n",
    "new_customers['gender'] = np.random.choice(['F','M','O'], 10) # random gender\n",
    "# One-hot encode gender:\n",
    "new_customers = new_customers.join(pd.get_dummies(new_customers['gender'])).drop('gender', axis=1) \n",
    "for column in ['M', 'F', 'O', 'U']:\n",
    "    new_customers = new_customers.rename(columns={column: (\"gender_\" + column.replace(\" \", \"_\"))})\n",
    "\n",
    "new_customers['bogo'] = 0\n",
    "new_customers['discount'] = 0\n",
    "new_customers['informational'] = 0\n",
    "\n",
    "new_customers['person'] = new_customers.index\n",
    "\n",
    "new_customers"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Assign all three offer types to each customer "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>age</th>\n",
       "      <th>income</th>\n",
       "      <th>gender_F</th>\n",
       "      <th>gender_M</th>\n",
       "      <th>gender_O</th>\n",
       "      <th>bogo</th>\n",
       "      <th>discount</th>\n",
       "      <th>informational</th>\n",
       "      <th>person</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>24</th>\n",
       "      <td>18</td>\n",
       "      <td>119267</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14</th>\n",
       "      <td>18</td>\n",
       "      <td>119267</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>18</td>\n",
       "      <td>119267</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>20</th>\n",
       "      <td>20</td>\n",
       "      <td>75724</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>20</td>\n",
       "      <td>75724</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>20</td>\n",
       "      <td>75724</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "    age  income  gender_F  gender_M  gender_O  bogo  discount  informational  \\\n",
       "24   18  119267         1         0         0     0         0              1   \n",
       "14   18  119267         1         0         0     0         1              0   \n",
       "4    18  119267         1         0         0     1         0              0   \n",
       "20   20   75724         0         0         1     0         0              1   \n",
       "10   20   75724         0         0         1     0         1              0   \n",
       "0    20   75724         0         0         1     1         0              0   \n",
       "\n",
       "    person  \n",
       "24       4  \n",
       "14       4  \n",
       "4        4  \n",
       "20       0  \n",
       "10       0  \n",
       "0        0  "
      ]
     },
     "execution_count": 41,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df1, df2, df3 = new_customers.copy(), new_customers.copy(), new_customers.copy()\n",
    "\n",
    "df1['bogo'] = 1\n",
    "df2['discount'] = 1\n",
    "df3['informational'] = 1\n",
    "\n",
    "frames = [df1, df2, df3]\n",
    "\n",
    "new_customers = pd.concat(frames, ignore_index=True)\n",
    "\n",
    "new_customers = new_customers.sort_values(by=['age', 'bogo', 'discount'])\n",
    "new_customers.head(6)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Deploy final XGBoost model into production"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "WARNING:sagemaker:Using already existing model: sagemaker-xgboost-200201-1703-001-0e60ab7e\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "---------------------!"
     ]
    }
   ],
   "source": [
    "xgb_predictor = xgb_attached.deploy(initial_instance_count=1, instance_type='ml.m4.xlarge')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Generate predictions for new customers"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>0</th>\n",
       "      <th>1</th>\n",
       "      <th>2</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0.258858</td>\n",
       "      <td>0.258858</td>\n",
       "      <td>0.536419</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0.251850</td>\n",
       "      <td>0.251850</td>\n",
       "      <td>0.534075</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0.109876</td>\n",
       "      <td>0.109876</td>\n",
       "      <td>0.615248</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0.193794</td>\n",
       "      <td>0.193794</td>\n",
       "      <td>0.613711</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0.155772</td>\n",
       "      <td>0.155772</td>\n",
       "      <td>0.598766</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "          0         1         2\n",
       "0  0.258858  0.258858  0.536419\n",
       "1  0.251850  0.251850  0.534075\n",
       "2  0.109876  0.109876  0.615248\n",
       "3  0.193794  0.193794  0.613711\n",
       "4  0.155772  0.155772  0.598766"
      ]
     },
     "execution_count": 43,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X = new_customers.drop(['person'], axis=1)\n",
    "X_id = new_customers['person']\n",
    "\n",
    "# We need to tell the endpoint what format the data we are sending is in\n",
    "xgb_predictor.content_type = 'text/csv'\n",
    "xgb_predictor.serializer = csv_serializer\n",
    "\n",
    "# Predictions is currently a comma delimited string:\n",
    "y_pred_raw = xgb_predictor.predict(X.values).decode('utf-8')\n",
    "\n",
    "# Convert predictions to a dataframe:\n",
    "y_pred = pd.DataFrame(y_pred_raw.split(\"],[\"))\n",
    "y_pred = y_pred[0].str.split(',', expand=True)\n",
    "\n",
    "# Fix the brackets and convert string to a float:\n",
    "y_pred[0] = y_pred[0].str.strip('[').astype(float)\n",
    "y_pred[1] = y_pred[0].astype(float)\n",
    "y_pred[2] = y_pred[2].str.strip(']').astype(float)\n",
    "\n",
    "y_pred.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Shut down the endpoint"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "metadata": {},
   "outputs": [],
   "source": [
    "xgb_predictor.delete_endpoint()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Explore prediction results\n",
    "\n",
    "#### Combine new customer data with the predictions"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>age</th>\n",
       "      <th>income</th>\n",
       "      <th>gender_F</th>\n",
       "      <th>gender_M</th>\n",
       "      <th>gender_O</th>\n",
       "      <th>bogo</th>\n",
       "      <th>discount</th>\n",
       "      <th>informational</th>\n",
       "      <th>person</th>\n",
       "      <th>0</th>\n",
       "      <th>1</th>\n",
       "      <th>2</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>20</td>\n",
       "      <td>75724</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.258858</td>\n",
       "      <td>0.258858</td>\n",
       "      <td>0.536419</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>24</td>\n",
       "      <td>118385</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0.251850</td>\n",
       "      <td>0.251850</td>\n",
       "      <td>0.534075</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>47</td>\n",
       "      <td>29441</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>2</td>\n",
       "      <td>0.109876</td>\n",
       "      <td>0.109876</td>\n",
       "      <td>0.615248</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>59</td>\n",
       "      <td>33835</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>3</td>\n",
       "      <td>0.193794</td>\n",
       "      <td>0.193794</td>\n",
       "      <td>0.613711</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>18</td>\n",
       "      <td>119267</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>4</td>\n",
       "      <td>0.155772</td>\n",
       "      <td>0.155772</td>\n",
       "      <td>0.598766</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   age  income  gender_F  gender_M  gender_O  bogo  discount  informational  \\\n",
       "0   20   75724         0         0         1     1         0              0   \n",
       "1   24  118385         0         0         1     1         0              0   \n",
       "2   47   29441         1         0         0     1         0              0   \n",
       "3   59   33835         0         1         0     1         0              0   \n",
       "4   18  119267         1         0         0     1         0              0   \n",
       "\n",
       "   person         0         1         2  \n",
       "0       0  0.258858  0.258858  0.536419  \n",
       "1       1  0.251850  0.251850  0.534075  \n",
       "2       2  0.109876  0.109876  0.615248  \n",
       "3       3  0.193794  0.193794  0.613711  \n",
       "4       4  0.155772  0.155772  0.598766  "
      ]
     },
     "execution_count": 45,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df = pd.concat([new_customers, y_pred], axis=1, sort=False)\n",
    "df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(30, 12)"
      ]
     },
     "execution_count": 46,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0    0.193794\n",
       "1    0.193794\n",
       "2    0.613711\n",
       "Name: 3, dtype: float64"
      ]
     },
     "execution_count": 47,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y_pred.iloc[3] # just to check if the merge was correct"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Improve data formatting"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>person</th>\n",
       "      <th>age</th>\n",
       "      <th>income</th>\n",
       "      <th>gender</th>\n",
       "      <th>offer</th>\n",
       "      <th>no view</th>\n",
       "      <th>no order</th>\n",
       "      <th>possible order</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1</td>\n",
       "      <td>24</td>\n",
       "      <td>118385</td>\n",
       "      <td>other</td>\n",
       "      <td>bogo</td>\n",
       "      <td>0.251850</td>\n",
       "      <td>0.251850</td>\n",
       "      <td>0.534075</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>2</td>\n",
       "      <td>47</td>\n",
       "      <td>29441</td>\n",
       "      <td>female</td>\n",
       "      <td>bogo</td>\n",
       "      <td>0.109876</td>\n",
       "      <td>0.109876</td>\n",
       "      <td>0.615248</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>6</td>\n",
       "      <td>32</td>\n",
       "      <td>70816</td>\n",
       "      <td>other</td>\n",
       "      <td>bogo</td>\n",
       "      <td>0.258077</td>\n",
       "      <td>0.258077</td>\n",
       "      <td>0.585960</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>27</th>\n",
       "      <td>7</td>\n",
       "      <td>66</td>\n",
       "      <td>47565</td>\n",
       "      <td>male</td>\n",
       "      <td>informational</td>\n",
       "      <td>0.291668</td>\n",
       "      <td>0.291668</td>\n",
       "      <td>0.496197</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>5</td>\n",
       "      <td>37</td>\n",
       "      <td>85146</td>\n",
       "      <td>female</td>\n",
       "      <td>bogo</td>\n",
       "      <td>0.091573</td>\n",
       "      <td>0.091573</td>\n",
       "      <td>0.618644</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "    person  age  income  gender          offer   no view  no order  \\\n",
       "1        1   24  118385   other           bogo  0.251850  0.251850   \n",
       "2        2   47   29441  female           bogo  0.109876  0.109876   \n",
       "6        6   32   70816   other           bogo  0.258077  0.258077   \n",
       "27       7   66   47565    male  informational  0.291668  0.291668   \n",
       "5        5   37   85146  female           bogo  0.091573  0.091573   \n",
       "\n",
       "    possible order  \n",
       "1         0.534075  \n",
       "2         0.615248  \n",
       "6         0.585960  \n",
       "27        0.496197  \n",
       "5         0.618644  "
      ]
     },
     "execution_count": 48,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Revert one-hot encoded columns back to categorical:\n",
    "df = df.rename(columns={\"gender_F\": \"female\",\n",
    "                        \"gender_M\": \"male\",\n",
    "                        \"gender_O\": \"other\"})\n",
    "gender = df[['female','male','other']]\n",
    "gender = pd.DataFrame(gender.idxmax(1))\n",
    "\n",
    "offer = df[['bogo','discount','informational']]\n",
    "offer = pd.DataFrame(offer.idxmax(1))\n",
    "\n",
    "df['gender'] = gender[0]\n",
    "df['offer'] = offer[0]\n",
    "\n",
    "# Rename columns:\n",
    "df = df.rename(columns={0: \"no view\",\n",
    "                        1: \"no order\",\n",
    "                        2: \"possible order\"})\n",
    "\n",
    "# Keep only these columns:\n",
    "columns_to_keep = ['person',\n",
    "                   'age',\n",
    "                   'income',\n",
    "                   'gender',\n",
    "                   'offer',\n",
    "                   'no view',\n",
    "                   'no order',\n",
    "                   'possible order']\n",
    "\n",
    "df = df[columns_to_keep]\n",
    "df.sample(5)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Explore the prediction results"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>person</th>\n",
       "      <th>age</th>\n",
       "      <th>income</th>\n",
       "      <th>gender</th>\n",
       "      <th>offer</th>\n",
       "      <th>no view</th>\n",
       "      <th>no order</th>\n",
       "      <th>possible order</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0</td>\n",
       "      <td>20</td>\n",
       "      <td>75724</td>\n",
       "      <td>other</td>\n",
       "      <td>bogo</td>\n",
       "      <td>0.258858</td>\n",
       "      <td>0.258858</td>\n",
       "      <td>0.536419</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>0</td>\n",
       "      <td>20</td>\n",
       "      <td>75724</td>\n",
       "      <td>other</td>\n",
       "      <td>discount</td>\n",
       "      <td>0.111028</td>\n",
       "      <td>0.111028</td>\n",
       "      <td>0.642354</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>20</th>\n",
       "      <td>0</td>\n",
       "      <td>20</td>\n",
       "      <td>75724</td>\n",
       "      <td>other</td>\n",
       "      <td>informational</td>\n",
       "      <td>0.093754</td>\n",
       "      <td>0.093754</td>\n",
       "      <td>0.511436</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "    person  age  income gender          offer   no view  no order  \\\n",
       "0        0   20   75724  other           bogo  0.258858  0.258858   \n",
       "10       0   20   75724  other       discount  0.111028  0.111028   \n",
       "20       0   20   75724  other  informational  0.093754  0.093754   \n",
       "\n",
       "    possible order  \n",
       "0         0.536419  \n",
       "10        0.642354  \n",
       "20        0.511436  "
      ]
     },
     "execution_count": 49,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "person = df[df.person == 0]\n",
    "person"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 50,
   "metadata": {},
   "outputs": [],
   "source": [
    "def plot_person_preds(df, i):\n",
    "    '''Plot predicted probability of belonging to each class'''\n",
    "    \n",
    "    df[df.person == i][[\"offer\",\n",
    "                        \"no view\",\n",
    "                        \"no order\",\n",
    "                        \"possible order\"]].plot(x=\"offer\", kind=\"barh\", stacked=True,\n",
    "                                                                    color=[custom_colors[1],\n",
    "                                                                           custom_colors[4],\n",
    "                                                                           custom_colors[0]])\n",
    "\n",
    "    plt.legend(loc=(1.05,0.75))\n",
    "    \n",
    "    plt.xlabel(\"predicted probability of belonging to each class\")\n",
    "    plt.ylabel(\"offer type\")\n",
    "\n",
    "    plt.show()\n",
    "    \n",
    "    return None"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 51,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Income: 75724 USD, other, 20 y/o:\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Income: 118385 USD, other, 24 y/o:\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Income: 29441 USD, female, 47 y/o:\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Income: 33835 USD, male, 59 y/o:\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Income: 119267 USD, female, 18 y/o:\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Income: 85146 USD, female, 37 y/o:\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Income: 70816 USD, other, 32 y/o:\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Income: 47565 USD, male, 66 y/o:\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Income: 115821 USD, male, 76 y/o:\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Income: 54680 USD, other, 24 y/o:\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "for i in range (10):\n",
    "    income = df.at[i, \"income\"]\n",
    "    gender = df.at[i, \"gender\"]\n",
    "    age = df.at[i, \"age\"]\n",
    "    print(\"\\nIncome: {} USD, {}, {} y/o:\".format(income, gender, age))\n",
    "    plot_person_preds(df, i)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "metadata": {},
   "outputs": [],
   "source": [
    "### FILE SAVING CHECKPOINT ###\n",
    "filename = 'data/new_customer_preds.csv'\n",
    "df.to_csv(filename)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "___\n",
    "Please refer to [PROJECT REPORT](https://evanca.github.io/machine-learning_optimizing-app-offers-with-starbucks/) for a conclusions and a high-level overview of the full project."
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.7.3"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}