experiment.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import os\n",
    "import pandas as pd\n",
    "from sklearn.linear_model import LogisticRegression\n",
    "from sklearn.model_selection import train_test_split,KFold\n",
    "import matplotlib.pylab as plt\n",
    "from sklearn.metrics import accuracy_score, confusion_matrix\n",
    "from utils import plot_confusion_matrix\n",
    "from sklearn.preprocessing import MinMaxScaler,StandardScaler\n",
    "from sklearn.svm import SVC\n",
    "from sklearn.ensemble import RandomForestClassifier\n",
    "from sklearn.neighbors import KNeighborsClassifier\n",
    "import matplotlib.pylab as plt\n",
    "from imblearn.over_sampling import SMOTE\n",
    "import sys"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# string to number"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "code_folding": []
   },
   "outputs": [],
   "source": [
    "col_names = [\"age\", \"job\", \"marital\", \"education\", \"default\",\"housing\",\"loan\",\"contact\",\"month\",\"day_of_week\"\n",
    "             ,\"duration\",\"campaign\",\"pdays\",\"previous\",\"poutcome\",\"emp.var.rate\",\"cons.price.idx\",\n",
    "             \"cons.conf.idx\",\"euribor3m\",\"nr.employed\",\"y\"]\n",
    "class_names = [\"yes\", \"no\"]\n",
    "job_to_int={\"admin.\":0,\"blue-collar\":1,\"entrepreneur\":2,\"housemaid\":3,\"management\":4,\"retired\":5,\"self-employed\":6,\n",
    "            \"services\":7,\"student\":8,\"technician\":9,\"unemployed\":10,\"unknown\":11}\n",
    "marital_to_int={\"divorced\":0,\"married\":1,\"single\":2,\"unknown\":3}\n",
    "education_to_int={\"basic.4y\":0,\"basic.6y\":1,\"basic.9y\":2,\"high.school\":3,\"illiterate\":4,\"professional.course\":5,\n",
    "                  \"university.degree\":6,\"unknown\":7}\n",
    "default_to_int={\"no\":0,\"yes\":1,\"unknown\":2}\n",
    "housing_to_int={\"no\":0,\"yes\":1,\"unknown\":2}\n",
    "loan_to_int={\"no\":0,\"yes\":1,\"unknown\":2}\n",
    "contact_to_int={\"cellular\":0,\"telephone\":1}\n",
    "month_to_int={\"jan\":0,\"feb\":1,\"mar\":2,\"apr\":3,\"may\":4,\"jun\":5,\"jul\":6,\n",
    "              \"aug\":7,\"sep\":8,\"oct\":9,\"nov\":10,\"dec\":11}\n",
    "day_to_int={\"mon\":0,\"tue\":1,\"wed\":2,\"thu\":3,\"fri\":4}\n",
    "poutcome_to_int={\"failure\":0,\"nonexistent\":1,\"success\":2}\n",
    "class_to_int = {\"yes\": 0, \"no\": 1}"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "data_dir = os.path.abspath(\"./data/\")\n",
    "data_file = os.path.abspath(data_dir + \"/blank.csv\")\n",
    "df = pd.read_csv(data_file,names=col_names,header=0,sep=';')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "   age        job  marital    education  default housing loan    contact  \\\n",
      "0   56  housemaid  married     basic.4y       no      no   no  telephone   \n",
      "1   57   services  married  high.school  unknown      no   no  telephone   \n",
      "2   37   services  married  high.school       no     yes   no  telephone   \n",
      "3   40     admin.  married     basic.6y       no      no   no  telephone   \n",
      "4   56   services  married  high.school       no      no  yes  telephone   \n",
      "\n",
      "  month day_of_week ...  campaign  pdays  previous     poutcome emp.var.rate  \\\n",
      "0   may         mon ...         1    999         0  nonexistent          1.1   \n",
      "1   may         mon ...         1    999         0  nonexistent          1.1   \n",
      "2   may         mon ...         1    999         0  nonexistent          1.1   \n",
      "3   may         mon ...         1    999         0  nonexistent          1.1   \n",
      "4   may         mon ...         1    999         0  nonexistent          1.1   \n",
      "\n",
      "   cons.price.idx  cons.conf.idx  euribor3m  nr.employed   y  \n",
      "0          93.994          -36.4      4.857       5191.0  no  \n",
      "1          93.994          -36.4      4.857       5191.0  no  \n",
      "2          93.994          -36.4      4.857       5191.0  no  \n",
      "3          93.994          -36.4      4.857       5191.0  no  \n",
      "4          93.994          -36.4      4.857       5191.0  no  \n",
      "\n",
      "[5 rows x 21 columns]\n"
     ]
    }
   ],
   "source": [
    "print(df.head())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "列名：\tage     \t , 有\t0    \t个Unknown数据；\n",
      "列名：\tjob     \t , 有\t330  \t个Unknown数据；\n",
      "列名：\tmarital \t , 有\t80   \t个Unknown数据；\n",
      "列名：\teducation\t , 有\t1731 \t个Unknown数据；\n",
      "列名：\tdefault \t , 有\t8597 \t个Unknown数据；\n",
      "列名：\thousing \t , 有\t990  \t个Unknown数据；\n",
      "列名：\tloan    \t , 有\t990  \t个Unknown数据；\n",
      "列名：\tcontact \t , 有\t0    \t个Unknown数据；\n",
      "列名：\tmonth   \t , 有\t0    \t个Unknown数据；\n",
      "列名：\tday_of_week\t , 有\t0    \t个Unknown数据；\n",
      "列名：\tduration\t , 有\t0    \t个Unknown数据；\n",
      "列名：\tcampaign\t , 有\t0    \t个Unknown数据；\n",
      "列名：\tpdays   \t , 有\t0    \t个Unknown数据；\n",
      "列名：\tprevious\t , 有\t0    \t个Unknown数据；\n",
      "列名：\tpoutcome\t , 有\t0    \t个Unknown数据；\n",
      "列名：\temp.var.rate\t , 有\t0    \t个Unknown数据；\n",
      "列名：\tcons.price.idx\t , 有\t0    \t个Unknown数据；\n",
      "列名：\tcons.conf.idx\t , 有\t0    \t个Unknown数据；\n",
      "列名：\teuribor3m\t , 有\t0    \t个Unknown数据；\n",
      "列名：\tnr.employed\t , 有\t0    \t个Unknown数据；\n",
      "列名：\ty       \t , 有\t0    \t个Unknown数据；\n"
     ]
    }
   ],
   "source": [
    "for i in df:\n",
    "    print(\"列名：\\t%-8s\\t , 有\\t%-5d\\t个Unknown数据；\"%(i,list(df[i]).count(\"unknown\")))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# create a new dataFrame"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "total=[]\n",
    "\n",
    "for index,row in df.iterrows():\n",
    "    temp = []\n",
    "    temp.append(row['age'])\n",
    "    temp.append(job_to_int[row['job']])\n",
    "    temp.append(marital_to_int[row['marital']])\n",
    "    temp.append(education_to_int[row['education']])\n",
    "    temp.append(default_to_int[row['default']])\n",
    "    temp.append(housing_to_int[row['housing']])\n",
    "    temp.append(loan_to_int[row['loan']])\n",
    "    temp.append(contact_to_int[row['contact']])\n",
    "    temp.append(month_to_int[row['month']])\n",
    "    temp.append(day_to_int[row['day_of_week']])\n",
    "    temp.append(row['duration'])\n",
    "    temp.append(row['campaign'])\n",
    "    temp.append(row['pdays'])\n",
    "    temp.append(row['previous'])\n",
    "    temp.append(poutcome_to_int[row['poutcome']])\n",
    "    temp.append(row['emp.var.rate'])\n",
    "    temp.append(row['cons.price.idx'])\n",
    "    temp.append(row['cons.conf.idx'])\n",
    "    temp.append(row['euribor3m'])\n",
    "    temp.append(row['nr.employed'])\n",
    "    temp.append(class_to_int[row['y']])\n",
    "    total.append(temp)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "               age          job       marital     education       default  \\\n",
      "count  41188.00000  41188.00000  41188.000000  41188.000000  41188.000000   \n",
      "mean      40.02406      3.72458      1.172769      3.747184      0.417525   \n",
      "std       10.42125      3.59456      0.608902      2.136482      0.812813   \n",
      "min       17.00000      0.00000      0.000000      0.000000      0.000000   \n",
      "25%       32.00000      0.00000      1.000000      2.000000      0.000000   \n",
      "50%       38.00000      2.00000      1.000000      3.000000      0.000000   \n",
      "75%       47.00000      7.00000      2.000000      6.000000      0.000000   \n",
      "max       98.00000     11.00000      3.000000      7.000000      2.000000   \n",
      "\n",
      "            housing          loan       contact         month   day_of_week  \\\n",
      "count  41188.000000  41188.000000  41188.000000  41188.000000  41188.000000   \n",
      "mean       0.571914      0.199767      0.365252      5.607896      1.979581   \n",
      "std        0.541209      0.456002      0.481507      2.040998      1.411514   \n",
      "min        0.000000      0.000000      0.000000      2.000000      0.000000   \n",
      "25%        0.000000      0.000000      0.000000      4.000000      1.000000   \n",
      "50%        1.000000      0.000000      0.000000      5.000000      2.000000   \n",
      "75%        1.000000      0.000000      1.000000      7.000000      3.000000   \n",
      "max        2.000000      2.000000      1.000000     11.000000      4.000000   \n",
      "\n",
      "           ...           campaign         pdays      previous      poutcome  \\\n",
      "count      ...       41188.000000  41188.000000  41188.000000  41188.000000   \n",
      "mean       ...           2.567593    962.475454      0.172963      0.930101   \n",
      "std        ...           2.770014    186.910907      0.494901      0.362886   \n",
      "min        ...           1.000000      0.000000      0.000000      0.000000   \n",
      "25%        ...           1.000000    999.000000      0.000000      1.000000   \n",
      "50%        ...           2.000000    999.000000      0.000000      1.000000   \n",
      "75%        ...           3.000000    999.000000      0.000000      1.000000   \n",
      "max        ...          56.000000    999.000000      7.000000      2.000000   \n",
      "\n",
      "       emp.var.rate  cons.price.idx  cons.conf.idx     euribor3m  \\\n",
      "count  41188.000000    41188.000000   41188.000000  41188.000000   \n",
      "mean       0.081886       93.575664     -40.502600      3.621291   \n",
      "std        1.570960        0.578840       4.628198      1.734447   \n",
      "min       -3.400000       92.201000     -50.800000      0.634000   \n",
      "25%       -1.800000       93.075000     -42.700000      1.344000   \n",
      "50%        1.100000       93.749000     -41.800000      4.857000   \n",
      "75%        1.400000       93.994000     -36.400000      4.961000   \n",
      "max        1.400000       94.767000     -26.900000      5.045000   \n",
      "\n",
      "        nr.employed             y  \n",
      "count  41188.000000  41188.000000  \n",
      "mean    5167.035911      0.887346  \n",
      "std       72.251528      0.316173  \n",
      "min     4963.600000      0.000000  \n",
      "25%     5099.100000      1.000000  \n",
      "50%     5191.000000      1.000000  \n",
      "75%     5228.100000      1.000000  \n",
      "max     5228.100000      1.000000  \n",
      "\n",
      "[8 rows x 21 columns]\n"
     ]
    }
   ],
   "source": [
    "new_df=pd.DataFrame(total)\n",
    "new_df.columns=col_names\n",
    "print (new_df.describe())"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# calculate the covariance matrix"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "age              -0.030399\n",
      "job              -0.025122\n",
      "marital          -0.046203\n",
      "education        -0.057799\n",
      "default           0.099324\n",
      "housing          -0.009552\n",
      "loan              0.005038\n",
      "contact           0.144773\n",
      "month            -0.037187\n",
      "day_of_week      -0.010051\n",
      "duration         -0.405274\n",
      "campaign          0.066357\n",
      "pdays             0.324914\n",
      "previous         -0.230181\n",
      "poutcome         -0.129789\n",
      "emp.var.rate      0.298334\n",
      "cons.price.idx    0.136211\n",
      "cons.conf.idx    -0.054878\n",
      "euribor3m         0.307771\n",
      "nr.employed       0.354678\n",
      "y                 1.000000\n",
      "Name: y, dtype: float64\n"
     ]
    }
   ],
   "source": [
    "corr=new_df.corr()\n",
    "print (corr.iloc[20,:])\n",
    "total = np.array(total)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# the most relevant to y ： duration\\nr.employed\n",
    "## Analytical balance"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "4640\n",
      "36548\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 504x504 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "\n",
    "len_yes=len(new_df[new_df['y']==1])\n",
    "len_no=len(new_df[new_df['y']==0])\n",
    "groups = new_df.groupby('y')\n",
    "fig, axes = plt.subplots(1, 1, figsize=(7, 7))\n",
    "\n",
    "for name, group in groups:\n",
    "    print (group['y'].size)\n",
    "    axes.scatter(group['duration'].values,group['nr.employed'], marker='o', label=name)\n",
    "\n",
    "axes.legend()\n",
    "plt.show()\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "36548 36548\n",
      "36548 4640\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "e:\\python3.7.1\\lib\\site-packages\\ipykernel_launcher.py:6: FutureWarning: Using a non-tuple sequence for multidimensional indexing is deprecated; use `arr[tuple(seq)]` instead of `arr[seq]`. In the future this will be interpreted as an array index, `arr[np.array(seq)]`, which will result either in an error or a different result.\n",
      "  \n",
      "e:\\python3.7.1\\lib\\site-packages\\ipykernel_launcher.py:7: FutureWarning: Using a non-tuple sequence for multidimensional indexing is deprecated; use `arr[tuple(seq)]` instead of `arr[seq]`. In the future this will be interpreted as an array index, `arr[np.array(seq)]`, which will result either in an error or a different result.\n",
      "  import sys\n",
      "e:\\python3.7.1\\lib\\site-packages\\matplotlib\\figure.py:98: MatplotlibDeprecationWarning: \n",
      "Adding an axes using the same arguments as a previous axes currently reuses the earlier instance.  In a future version, a new instance will always be created and returned.  Meanwhile, this warning can be suppressed, and the future behavior ensured, by passing a unique label to each axes instance.\n",
      "  \"Adding an axes using the same arguments as a previous axes \"\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "X=total[:,0:20]\n",
    "y = df['y'].values\n",
    "y = np.array([class_to_int[c] for c in y])\n",
    "more_sample = SMOTE(random_state=21)\n",
    "X, y = more_sample.fit_sample(X, y)\n",
    "len_new_yes=len(y[[y==1]])\n",
    "len_new_no=len(y[[y==0]])\n",
    "print (len_new_yes,len_new_no)\n",
    "print (len_yes,len_no)\n",
    "\n",
    "plt.xlabel('Class for y')\n",
    "plt.ylabel('Number')\n",
    "plt.title('The statistics of y  dataset')\n",
    "x_new=[0.01,1.01]\n",
    "x_old=[-0.01,0.99]\n",
    "old=[len_no,len_yes]\n",
    "new=[len_new_no,len_new_yes]\n",
    "plt.subplot(1,1,1)\n",
    "plt.ylim=(0, 10)\n",
    "plt.bar(x_new, new, facecolor='red', width=0.01, label = 'new')\n",
    "plt.bar(x_old, old, facecolor='blue', width=0.01, label = 'old')\n",
    "plt.legend()\n",
    "\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Normalized data\n",
    "\n",
    "### 10、19\n",
    "\n",
    "#### delete 5\\6 because of low covariance"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[[ 261.         5191.        ]\n",
      " [ 149.         5191.        ]\n",
      " [ 226.         5191.        ]\n",
      " ...\n",
      " [ 816.44274115 5228.1       ]\n",
      " [ 193.06729218 5017.5       ]\n",
      " [ 175.7696771  5076.2       ]]\n",
      "[[3.         0.85973535]\n",
      " [7.         0.85973535]\n",
      " [7.         0.85973535]\n",
      " ...\n",
      " [2.21370573 1.        ]\n",
      " [0.         0.20378072]\n",
      " [0.2303229  0.42570888]]\n"
     ]
    }
   ],
   "source": [
    "print (X[:,[10,19]])\n",
    "scaler = MinMaxScaler( )\n",
    "scaler.fit(X[:,10].reshape(-1,1))\n",
    "X[:,10]=scaler.transform(X[:,1].reshape(-1,1)).reshape(1,-1)\n",
    "scal = MinMaxScaler( )\n",
    "scal.fit(X[:,19].reshape(-1,1))\n",
    "X[:,19]=scal.transform(X[:,19].reshape(-1,1)).reshape(1,-1)\n",
    "print (X[:,[1,19]])\n",
    "\n",
    "X = X[:,[0,1,2,3,4,7,8,9,10,11,12,13,14,15,16,17,18,19]]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [],
   "source": [
    "def draw_acc(x_plot,y_plot):\n",
    "    fig, ax = plt.subplots()\n",
    "    ax.plot(x_plot, y_plot, 'o')\n",
    "    plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# logistic regression\n",
    "## 25% test     75%train\n",
    "### find the best C  on the same data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "e:\\python3.7.1\\lib\\site-packages\\sklearn\\linear_model\\logistic.py:758: ConvergenceWarning: lbfgs failed to converge. Increase the number of iterations.\n",
      "  \"of iterations.\", ConvergenceWarning)\n",
      "e:\\python3.7.1\\lib\\site-packages\\sklearn\\linear_model\\logistic.py:758: ConvergenceWarning: lbfgs failed to converge. Increase the number of iterations.\n",
      "  \"of iterations.\", ConvergenceWarning)\n",
      "e:\\python3.7.1\\lib\\site-packages\\sklearn\\linear_model\\logistic.py:758: ConvergenceWarning: lbfgs failed to converge. Increase the number of iterations.\n",
      "  \"of iterations.\", ConvergenceWarning)\n",
      "e:\\python3.7.1\\lib\\site-packages\\sklearn\\linear_model\\logistic.py:758: ConvergenceWarning: lbfgs failed to converge. Increase the number of iterations.\n",
      "  \"of iterations.\", ConvergenceWarning)\n",
      "e:\\python3.7.1\\lib\\site-packages\\sklearn\\linear_model\\logistic.py:758: ConvergenceWarning: lbfgs failed to converge. Increase the number of iterations.\n",
      "  \"of iterations.\", ConvergenceWarning)\n",
      "e:\\python3.7.1\\lib\\site-packages\\sklearn\\linear_model\\logistic.py:758: ConvergenceWarning: lbfgs failed to converge. Increase the number of iterations.\n",
      "  \"of iterations.\", ConvergenceWarning)\n",
      "e:\\python3.7.1\\lib\\site-packages\\sklearn\\linear_model\\logistic.py:758: ConvergenceWarning: lbfgs failed to converge. Increase the number of iterations.\n",
      "  \"of iterations.\", ConvergenceWarning)\n",
      "e:\\python3.7.1\\lib\\site-packages\\sklearn\\linear_model\\logistic.py:758: ConvergenceWarning: lbfgs failed to converge. Increase the number of iterations.\n",
      "  \"of iterations.\", ConvergenceWarning)\n",
      "e:\\python3.7.1\\lib\\site-packages\\sklearn\\linear_model\\logistic.py:758: ConvergenceWarning: lbfgs failed to converge. Increase the number of iterations.\n",
      "  \"of iterations.\", ConvergenceWarning)\n",
      "e:\\python3.7.1\\lib\\site-packages\\sklearn\\linear_model\\logistic.py:758: ConvergenceWarning: lbfgs failed to converge. Increase the number of iterations.\n",
      "  \"of iterations.\", ConvergenceWarning)\n",
      "e:\\python3.7.1\\lib\\site-packages\\sklearn\\linear_model\\logistic.py:758: ConvergenceWarning: lbfgs failed to converge. Increase the number of iterations.\n",
      "  \"of iterations.\", ConvergenceWarning)\n",
      "e:\\python3.7.1\\lib\\site-packages\\sklearn\\linear_model\\logistic.py:758: ConvergenceWarning: lbfgs failed to converge. Increase the number of iterations.\n",
      "  \"of iterations.\", ConvergenceWarning)\n",
      "e:\\python3.7.1\\lib\\site-packages\\sklearn\\linear_model\\logistic.py:758: ConvergenceWarning: lbfgs failed to converge. Increase the number of iterations.\n",
      "  \"of iterations.\", ConvergenceWarning)\n",
      "e:\\python3.7.1\\lib\\site-packages\\sklearn\\linear_model\\logistic.py:758: ConvergenceWarning: lbfgs failed to converge. Increase the number of iterations.\n",
      "  \"of iterations.\", ConvergenceWarning)\n",
      "e:\\python3.7.1\\lib\\site-packages\\sklearn\\linear_model\\logistic.py:758: ConvergenceWarning: lbfgs failed to converge. Increase the number of iterations.\n",
      "  \"of iterations.\", ConvergenceWarning)\n",
      "e:\\python3.7.1\\lib\\site-packages\\sklearn\\linear_model\\logistic.py:758: ConvergenceWarning: lbfgs failed to converge. Increase the number of iterations.\n",
      "  \"of iterations.\", ConvergenceWarning)\n",
      "e:\\python3.7.1\\lib\\site-packages\\sklearn\\linear_model\\logistic.py:758: ConvergenceWarning: lbfgs failed to converge. Increase the number of iterations.\n",
      "  \"of iterations.\", ConvergenceWarning)\n",
      "e:\\python3.7.1\\lib\\site-packages\\sklearn\\linear_model\\logistic.py:758: ConvergenceWarning: lbfgs failed to converge. Increase the number of iterations.\n",
      "  \"of iterations.\", ConvergenceWarning)\n",
      "e:\\python3.7.1\\lib\\site-packages\\sklearn\\linear_model\\logistic.py:758: ConvergenceWarning: lbfgs failed to converge. Increase the number of iterations.\n",
      "  \"of iterations.\", ConvergenceWarning)\n",
      "e:\\python3.7.1\\lib\\site-packages\\sklearn\\linear_model\\logistic.py:758: ConvergenceWarning: lbfgs failed to converge. Increase the number of iterations.\n",
      "  \"of iterations.\", ConvergenceWarning)\n",
      "e:\\python3.7.1\\lib\\site-packages\\sklearn\\linear_model\\logistic.py:758: ConvergenceWarning: lbfgs failed to converge. Increase the number of iterations.\n",
      "  \"of iterations.\", ConvergenceWarning)\n",
      "e:\\python3.7.1\\lib\\site-packages\\sklearn\\linear_model\\logistic.py:758: ConvergenceWarning: lbfgs failed to converge. Increase the number of iterations.\n",
      "  \"of iterations.\", ConvergenceWarning)\n",
      "e:\\python3.7.1\\lib\\site-packages\\sklearn\\linear_model\\logistic.py:758: ConvergenceWarning: lbfgs failed to converge. Increase the number of iterations.\n",
      "  \"of iterations.\", ConvergenceWarning)\n",
      "e:\\python3.7.1\\lib\\site-packages\\sklearn\\linear_model\\logistic.py:758: ConvergenceWarning: lbfgs failed to converge. Increase the number of iterations.\n",
      "  \"of iterations.\", ConvergenceWarning)\n",
      "e:\\python3.7.1\\lib\\site-packages\\sklearn\\linear_model\\logistic.py:758: ConvergenceWarning: lbfgs failed to converge. Increase the number of iterations.\n",
      "  \"of iterations.\", ConvergenceWarning)\n",
      "e:\\python3.7.1\\lib\\site-packages\\sklearn\\linear_model\\logistic.py:758: ConvergenceWarning: lbfgs failed to converge. Increase the number of iterations.\n",
      "  \"of iterations.\", ConvergenceWarning)\n",
      "e:\\python3.7.1\\lib\\site-packages\\sklearn\\linear_model\\logistic.py:758: ConvergenceWarning: lbfgs failed to converge. Increase the number of iterations.\n",
      "  \"of iterations.\", ConvergenceWarning)\n",
      "e:\\python3.7.1\\lib\\site-packages\\sklearn\\linear_model\\logistic.py:758: ConvergenceWarning: lbfgs failed to converge. Increase the number of iterations.\n",
      "  \"of iterations.\", ConvergenceWarning)\n",
      "e:\\python3.7.1\\lib\\site-packages\\sklearn\\linear_model\\logistic.py:758: ConvergenceWarning: lbfgs failed to converge. Increase the number of iterations.\n",
      "  \"of iterations.\", ConvergenceWarning)\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.25, random_state=42)\n",
    "x_plot=[]\n",
    "y_plot=[]\n",
    "for step in range(1, 30):\n",
    "    logistic_regression=LogisticRegression(C=step/100., solver='sag', multi_class='multinomial')\n",
    "    logistic_regression.fit(X_train, y_train)\n",
    "    pred=logistic_regression.predict(X_test)\n",
    "    acc=accuracy_score(y_test, pred)\n",
    "    x_plot.append(step)\n",
    "    y_plot.append(acc)\n",
    "\n",
    "draw_acc(x_plot=x_plot,y_plot=y_plot)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# C=0.19 ensemble"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "code_folding": []
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "e:\\python3.7.1\\lib\\site-packages\\sklearn\\linear_model\\sag.py:334: ConvergenceWarning: The max_iter was reached which means the coef_ did not converge\n",
      "  \"the coef_ did not converge\", ConvergenceWarning)\n",
      "e:\\python3.7.1\\lib\\site-packages\\sklearn\\linear_model\\sag.py:334: ConvergenceWarning: The max_iter was reached which means the coef_ did not converge\n",
      "  \"the coef_ did not converge\", ConvergenceWarning)\n",
      "e:\\python3.7.1\\lib\\site-packages\\sklearn\\linear_model\\sag.py:334: ConvergenceWarning: The max_iter was reached which means the coef_ did not converge\n",
      "  \"the coef_ did not converge\", ConvergenceWarning)\n",
      "e:\\python3.7.1\\lib\\site-packages\\sklearn\\linear_model\\sag.py:334: ConvergenceWarning: The max_iter was reached which means the coef_ did not converge\n",
      "  \"the coef_ did not converge\", ConvergenceWarning)\n",
      "e:\\python3.7.1\\lib\\site-packages\\sklearn\\linear_model\\sag.py:334: ConvergenceWarning: The max_iter was reached which means the coef_ did not converge\n",
      "  \"the coef_ did not converge\", ConvergenceWarning)\n",
      "e:\\python3.7.1\\lib\\site-packages\\sklearn\\linear_model\\sag.py:334: ConvergenceWarning: The max_iter was reached which means the coef_ did not converge\n",
      "  \"the coef_ did not converge\", ConvergenceWarning)\n",
      "e:\\python3.7.1\\lib\\site-packages\\sklearn\\linear_model\\sag.py:334: ConvergenceWarning: The max_iter was reached which means the coef_ did not converge\n",
      "  \"the coef_ did not converge\", ConvergenceWarning)\n",
      "e:\\python3.7.1\\lib\\site-packages\\sklearn\\linear_model\\sag.py:334: ConvergenceWarning: The max_iter was reached which means the coef_ did not converge\n",
      "  \"the coef_ did not converge\", ConvergenceWarning)\n",
      "e:\\python3.7.1\\lib\\site-packages\\sklearn\\linear_model\\sag.py:334: ConvergenceWarning: The max_iter was reached which means the coef_ did not converge\n",
      "  \"the coef_ did not converge\", ConvergenceWarning)\n"
     ]
    }
   ],
   "source": [
    "model=[]\n",
    "res=[]\n",
    "for i in range(0, 15):\n",
    "    X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.25, random_state=i)\n",
    "    logistic_regression=LogisticRegression(C=0.19, solver='sag', multi_class='multinomial')\n",
    "    logistic_regression.fit(X_train, y_train)\n",
    "    pred=logistic_regression.predict(X_test)\n",
    "    acc=accuracy_score(y_test, pred)\n",
    "    model.append(logistic_regression)\n",
    "    res.append(acc)\n",
    "\n",
    "avg_acc=sum(res)/15.\n",
    "print ('final result : %f ' %avg_acc)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Confusion matrix, without normalization\n",
      "[[6781 2410]\n",
      " [2673 6410]]\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAUYAAAEYCAYAAAAgU193AAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDMuMC4yLCBodHRwOi8vbWF0cGxvdGxpYi5vcmcvOIA7rQAAIABJREFUeJzt3XecVNX5x/HPd1mqgKJiAVRQELCBNEs0IdiwRCU2rFiiscYSW6KJlWiMxvJTY+zGEsQWsUREjUYNKKJYUERQlF6kCKLU5/fHObPMjjuzs+vdnd3Z583rvpi599xzz92Zffbcdh6ZGc4559YqKXQDnHOurvHA6JxzGTwwOudcBg+MzjmXwQOjc85l8MDonHMZGnRglNRc0jOSFkt67EfUc7SkF5NsW6FI2l3Sp3Vle5I6SjJJpbXVppoi6XJJD8XXm0taKqlRwtuo1c+vWNWLwCjpKEnvxC/SLEn/lrRbAlUfCmwMbGBmh1W3EjN72Mz2TqA9NSoGmM65ypjZ62bWtbbalLk9SVMl7Vlb2y8UM/vKzFqa2eofU0/mZ1rbn1+xqvOBUdJ5wE3AnwhBbHPgduCgBKrfAphkZqsSqKveK4ZeWVIU1PnfD1dDzKzOTsC6wFLgsBxlmhIC58w43QQ0jcv6A9OB3wJzgVnACXHZFcAKYGXcxknA5cBDaXV3BAwoje+PBz4HlgBfAEenzX8jbb1dgbHA4vj/rmnLXgWuAt6M9bwIbJhl31LtvzCt/QcD+wGTgAXA79PK9wNGA4ti2VuBJnHZf+O+fBv394i0+i8CZgMPpubFdbaK2+gV37cD5gP98/jsHgB+G1+3j9s+Pb7vHOtVxvYeBNYA38U2Xpj2GQwBvorbvyTHdu8HbgOeiz/ft4CtqvDZDI2fzXexna8CVwP/i216BtgAeBj4JtbRMa2Om4Fpcdk4YPe0ZZcTv19p+1UK7BLrTk3fA1N/zGeats3ucR8WAROAA/P9WTXkqeANqOSXayCwihiYspS5EhgDbAS0jV/gq+Ky/nH9K4HGhICyDGiT+UXN8j79y7tO/LJ3jcs2BbaNr48nBkZgfWAhcGxc78j4foO4/FVgCrA10Dy+vzbLvqXa/8fY/pOBecAjQCtg2/hLtGUs3xvYOW63I/AJcE5afQZ0rqD+PxP+wDSv4Bfr5FhPC2AkcH2en92JwDPx9VFxnx9NW/Z0WhvStzcV2LOCz+Cu2L4ewHKge5bt3k8Iuv3iz+FhYFgVPpuv4s+1NP7MXwUmE/5IrAt8TPijtGcs8w/gvrTtH0MInKWEP8izgWaZ3y8y/uimrZ/a5jU/4jOdnlbXZOD3QBNgACEAdq3sZ9XQp7p+qLABMN9yH+oeDVxpZnPNbB6hJ3hs2vKVcflKM3ue8Je1uudg1gDbSWpuZrPMbEIFZfYHPjOzB81slZn9E5gI/CKtzH1mNsnMvgOGAz1zbHMlMNTMVgLDgA2Bm81sSdz+BGAHADMbZ2Zj4nanAn8HfpbHPl1mZstje8oxs7uAzwi9iU2BSyqpL+U1YPd4OPpT4DrgJ3HZz+LyqrjCzL4zs/eB9wkBMpsnzezt+L15mLU/33w+m/vNbEJcvjLOu8/MppjZYuDfwBQzeynW/xiwY2plM3vIzL6O699A+INTle/bLYQe4CWxvup8pik7Ay0Jf3hXmNkrwLOEPwgp2X5WDVpdD4xfAxtWcu6rHfBl2vsv47yyOjIC6zLCl6VKzOxbwqHKqcAsSc9J6pZHe1Jtap/2fnYV2vO1rT1Bnwpcc9KWf5daX9LWkp6VNFvSN4TzshvmqBtgnpl9X0mZu4DtgP8zs+WVlAXAzKYQ/gj1BHYn/ELOlNSV6gXGqvzMspXN57OZVkF9mT/vCn/+AJJ+K+mTeKfDIkIvs7LPILXurwk9vqPMbE2cV53PNKUdMC1VV/RjvosNRl0PjKMJh4oH5ygzk3ARJWXzOK86viUcMqZskr7QzEaa2V6EntNEQsCorD2pNs2oZpuq4m+EdnUxs9aEQyhVsk7O4ZUktSSct70HuFzS+lVoz2uEK/9NzGxGfH8c0AYYX532/Ej5fDbV3r6k3Qnnaw8nnK5Zj3Aus7LPILXuVcBBsWeaUp3PNGUmsFnGRaTa+i7Wa3U6MMYvyB+B2yQdLKmFpMaS9pV0XSz2T+BSSW0lbRjLP1TNTY4HfhrvMVsX+F1qgaSNJR0oaR3COa6lQEW3WjwPbB1vMSqVdASwDaHHVNNaEc6DLo292dMyls8BtqxinTcD48zsV4ST9HekFsT78l7Nse5rwJmEiwQQzp2dRTgfm+02leq0MV81/dm0IpyznQeUSvoj0LqylSRtBjwKHGdmkyqos7qf6VuEP/YXxt+b/oTTBsPy252Gq04HRgAz+ytwHnAp4Qs3jfDL9q9Y5GrgHeAD4EPg3TivOtsaRfiCfkC4opj+C1NCOJk+k3DC+mfA6RXU8TVwQCz7NeHK6gFmNr86baqi8wkXOpYQerOPZiy/HHhA0iJJh1dWmaSDCBfATo2zzgN6STo6vt+McAU3m9cIv9ipwPgGoUf+36xrwDWEP3SLJJ1fWRurohY+m5GEc5CTCIes31PxoXmmPQhHJ4/He3WXSkqdv672Z2pmK4ADgX0JV/NvJwTfidXYtwZFZj5QraseSeOBPWLAca5oeGB0zrkMdf5Q2jnnapsHRuecy+CB0TnnMtTLQQNU2tzUpFWhm+Fy2LH75oVugsvDu++Om29mbZOqr1HrLcxW/eABqgrZd/NGmtnApLadpPoZGJu0omnXSu82cQX05lu3FroJLg/NGyvzSaAfxVZ9l/fv5vfjb8v3CZ5aVy8Do3OurhIUwWhtHhidc8kRUJLooOQF4YHROZcs5fsod93lgdE5lyA/lHbOuR/yHqNzzqWR/Byjc879gB9KO+dcBj+Uds65dMVx8aX+74Fzru4QoceYz1RZVdJ6kh6XNDHm0dkljho/Q9L4OO2XVv53kiZL+lTSPmnzB8Z5kyVdnM9ueI/ROZcgQUliYeVm4AUzO1RSE8Lo7/sAN5rZ9eW2Km0DDCakvm0HvCRp67j4NmAvQg71sZJGmNnHuTbsgdE5l6ySH3+OUVJrQtrd46EsTcMKZe9pHkTIib0c+ELSZEK+bIDJZvZ5rHdYLJszMPqhtHMuOSKcY8xnCqmR30mbTkmraUtCjqf7JL0n6e6YiA7gTEkfSLpXUps4rz3l8+tMj/Oyzc/JA6NzLln5n2Ocb2Z90qY702opBXoBfzOzHQnZDi8mpJPdipCvfBZwQ2qrFbTEcszPyQOjcy5B8QbvfKbcpgPTzeyt+P5xoJeZzTGz1Wa2hpA1sV9a+c3S1u9AyOiZbX5OHhidc8nK/1A6KzObDUyT1DXO2gP4WNKmacUGAR/F1yOAwZKaSuoEdAHeBsYCXSR1ihdwBseyOfnFF+dccvK8FSdPZwEPx4D2OXACcIuknoTD4anArwHMbIKk4YSLKquAM8xsdWiSziTk/G4E3GtmEzI3lMkDo3MuWQnd4G1m44E+GbOPzVF+KDC0gvnPA89XZdseGJ1zyfJHAp1zLp2PruOcc+Wl7mOs5zwwOucSVByDSHhgdM4ly88xOudcBu8xOudcGk9t4JxzFfBDaeecKy/H0GD1hgdG51xiwgDeHhidc24tCSUwUG2heWB0ziXKe4zOOZfBA6NzzmXwwOicc+lExckE6pn6f4u6c67OEKKkpCSvqdK6Ks4rvb6kUZI+i/+3iWUl6ZaYO/oDSb3S6hkSy38maUg+++GB0TmXKEl5TXlI5ZXuBvQAPiEkxHrZzLoAL8f3APsS0hl0AU4hJM1C0vrAZcBOhPwwl6VlFszKA6NzLlFJBMa0vNL3QMgrbWaLCDmhH4jFHgAOjq8PAv5hwRhgvZgfZh9glJktMLOFwChgYGX74IHROZccVWGqXl7pjc1sFkD8f6NYPtG80n7xxTmXmNQ5xjzNN7PMnC4pqbzSZ5nZW5JuZu1hc8Wb/iHPK+2cqxsSOsdYYV5pYE4qhWr8f25aec8r7Zyro/I/lM4qW15pQk7o1JXlIcDT8fUI4Lh4dXpnYHE81B4J7C2pTbzosnecl5MfSjvnkqNEb/CuKK90CTBc0knAV8BhsezzwH7AZGBZLIuZLZB0FTA2lrvSzBZUtmEPjM65RCUVGLPklYbQe8wsa8AZWeq5F7i3Ktv2wOicS0wVL77UWR4YnXPJ8kcCXS7rtmzOI385ifFPXsp7T1zKTjt04sFrT2DMsIsZM+xiJj53BWOGhTsQSktLuOvKYxk7/Pe898SlnH/i3mX13HHZ0Xz58jW889jvC7UrRWnatGnss+fP6bl9d3r12JZbb7m53PIb/3o9zRuL+fPnA/DpxIn8bLddWHedptz41+vLlX1x5AvssG1Xtu3Wmb9cd22t7UOdo0SffCkY7zHWoOsvPJQX//cxR11wD41LG9GiWROOvfi+suXXnjeIxUu/A+CQPXvRtEkpfQ//E82bNea9Jy5l+L/f4atZC3jwmTHc8ehr3H3VcYXalaJUWlrKtdfdwI69erFkyRJ23ak3e+y5F9232YZp06bxykuj2GzzzcvKt1l/fW648RaeGfGvcvWsXr2ac35zBs/9exTtO3Rgt537csABB9J9m21qe5fqhLoe9PLhPcYa0mqdZuzWayvuf2o0ACtXrS4LgimH7NWL4S+MA8AwWjRrQqNGJTRv2oQVK1ez5NvvAXjz3SksWLysdnegAdh0003ZsVcYa6BVq1Z069admTNnAHDh+ecy9Jrryv2Sb7TRRvTp25fGjRuXq2fs22+z1Vad6bTlljRp0oTDjhjMs888TUPlPUaXVaf2GzB/4VLuvOIYtt+6Pe99Mo3zr3ucZd+vAOAnvbZizoIlTPlqHgBPvvQeB/TfgS9GDaVFsyZceP2TLPzGg2Ft+XLqVMaPf4++/Xbi2WdG0K5de3bo0SOvdWfOnEGHDmvvIW7fvgNvv/1WjjWKWzGkNvAeYw0pLW1Ez26bcddjr7PLkX9m2XfLOf/EvcqWHz6wD4+98E7Z+77bdmT16jVsufcldN//Ms4+dgAd229QiKY3OEuXLuXIww/hLzfcRGlpKX++Zih/vPzKvNcPd4qUV9d7RDUl395iXf/5eGCsITPmLGTG3EWM/ehLAJ56aTw9u4VeRaNGJRw0oAePj3y3rPzh+/bhxf99zKpVa5i3cCmjx39O7202r7Bul5yVK1dy5OGHcMSRR3PwoF/y+ZQpfDn1C/r17kHXzh2ZMX06u/TrxezZs7PW0b59B6ZPXztOwYwZ02nXrl1tNL9O8sCYhaSrJJ2d9n6opN9IukDS2DiQ5BVx2TqSnpP0vqSPJB1RE22qbXO+XsL02QvpskUY/KN/v65M/Dz8cg3YqSuTps5hxtxFZeWnz15A/77h6acWzZrQb4eOfDp1Tu03vAExM049+SS6duvO2eeeB8B222/PVzPn8unkqXw6eSrtO3Rg9Nvvsskmm2Stp0/fvkye/BlTv/iCFStW8Nijw9j/gANrazfqnGIIjDV1jvEe4EngZkklwGDg94Q71vsR7nQaIemnQFtgppntDyBp3YoqjEMShWGJGresoWYn67w/P8Z9fzqeJqWNmDpjPqdc9hAAh+3Tu+yiS8odj/6XO684hnGPX4IEDz49ho8+C8+6P3DN8ezeuwsbrteSyS9cxVV3PM8D/xpd6/tTbP735ps88vCDbLfd9uzUuycAV1z9Jwbuu1+F5WfPns1Pdu7Dkm++oaSkhFtvuYn3PviY1q1bc+PNt/KL/fdh9erVDDn+RLbZdtva3JU6pRjOMaqi8yOJVCyNAi4ENgZ+BUwFDgVS3aSWwDXA64SHuocDz5rZ65XVXdJiI2va9fAaaLVLysKxtxa6CS4PzRtrXI6hv6qs6SZdrMPRt+RV9vO/7pfotpNUk1el7waOBzYhPKe4B3CNmf09s6Ck3oQHwK+R9KKZ5X/m2zlXZwio40fJeanJwPgUcCXQGDgKWAVcJelhM1sqqT2wMrZhgZk9JGkpIZg65+qlun/+MB81FhjNbIWk/wCLzGw18KKk7sDo+INbChwDdAb+ImkNIVCeVlNtcs7VvCKIizUXGONFl51ZO14aZnYzIfNXuinkMXCkc64eEJQUwcWXmrpdZxvCgJEvm9lnNbEN51zdI0JgzGeqtC5pqqQPJY2X9E6cd7mkGXHeeEn7pZX/nUJe6U8l7ZM2f2CcN1lSrrwxZWqkx2hmHxOyfDnnGpiED6V/bmbzM+bdaGblhjeKnbHBwLZAO+AlSVvHxbcBexHyv4yVNCLGqKz8WWnnXKIKdPHlIGCYmS0HvpA0mXDPNMBkM/s8tm1YLJszMPojgc65xEhVOpTOlVcaQprTFyWNy1h2Znx67t6Y4Ao8r7Rzru6q0u06ufJKA/zEzGZK2ggYJWki8DfgKkLQvAq4ATiR7PmjK+r8eV5p51ztkvKbKmNmM+P/cwn3RfczszlmttrM1gB3sfZw2fNKO+fqriQGkYiDy7RKvSbkg/5I0qZpxQYBH8XXI4DBkppK6gR0Ad4mpE3tIqmTQhrWwbFsTn4o7ZxLTp69wTxsDDwVA2gp8IiZvSDpQUk9CYfDU4FfA5jZBEnDCRdVVgFnxAdLkHQm4V7pRsC9Zjahso17YHTOJSZ1H+OPFa8i/2AIdTM7Nsc6Q4GhFcx/Hni+Ktv3wOicS5Q/K+2ccxmKIC56YHTOJUjeY3TOuXJEfs9B13UeGJ1ziSqCDqMHRudcsvxQ2jnn0iV3H2NBeWB0ziUm5Hyp/5HRA6NzLlF+8cU55zJ4j9E559L5OUbnnCtPnj7VOed+qAjiogdG51yyGhXzxRdJrXOtaGbfJN8c51x9pgbwrPQEwmCQ6XuZem/A5jXYLudcPZVUh1HSVGAJsBpYZWZ9JK0PPAp0JAxUe7iZLVSIxjcD+wHLgOPN7N1YzxDg0ljt1Wb2QGXbzhoYzWyzbMuccy6bhHuMmXmlLwZeNrNrJV0c318E7EtIZ9AF2ImQNGunGEgvA/oQOnTjYl7phbk2mlfOF0mDJf0+vu4gqXfV9s051xAIKJHymqrpICDV43sAODht/j8sGAOsF/PD7AOMMrMFMRiOAgZWtpFKA6OkW4GfA6khxZcBd1RlT5xzDUeJ8puoXl7pjc1sFkD8f6M4v9bzSu9qZr0kvRcbsyBm23LOufLyyACYpjp5pbNuuYJ5mddI0ufnlM+h9EpJJanKJG0ArMljPedcA1STeaWBOakUqvH/ubF4reeVvg14Amgr6QrgDeDPeaznnGtgkjrHmC2vNCEn9JBYbAjwdHw9AjhOwc7A4nioPRLYW1IbSW1iPSMr249KD6XN7B+SxgF7xlmHmdlHudZxzjVcCY2uky2v9FhguKSTgK+Aw2L55wm36kwmXAc5AcpO/V0FjI3lrjSzBZVtPN8nXxoBKwmH03ldyXbONTz5HiZXJkde6a+BPSqYb8AZWeq6F7i3KtvP56r0JcA/gXaE4/NHJP2uKhtxzjUcNXy7Tq3Ip8d4DNDbzJYBSBoKjAOuqcmGOefqp7od8vKTT2D8MqNcKfB5zTTHOVefieIfROJGwjnFZcAESSPj+70JV6adc668qt3HWGfl6jGmrjxPAJ5Lmz+m5prjnKvviiAu5hxE4p7abIhzrjgUe48RAElbAUOBbYBmqflmtnUNtss5Vw+FG7wL3YofL597Eu8H7iPs877AcGBYDbbJOVePFcPtOvkExhZmNhLAzKaY2aWE0Xacc64cqTgCYz636yyPo+NOkXQqMIO1Q/0451w5dTzm5SWfwHgu0BL4DeFc47rAiTXZKOdc/dUgLr6Y2Vvx5RLWDlbrnHMVKoK4mPMG76fIMaCjmf2yRlrknKu3JBX3ky/ArbXWiiraodtmvPTfmwrdDJdDm/2uL3QTXIEU9aG0mb1cmw1xzhWHYhiXsBj2wTlXR4jQY8xnyqs+qZGk9yQ9G9/fL+kLSePj1DPOl6RbJE2W9IGkXml1DJH0WZyGZNtWunwHqnXOubyUJtvdOhv4BGidNu8CM3s8o1zt55UGkNQ037LOuYYpjOCdTI9RUgdgf+DuPDZd63ml+0n6EPgsvu8h6f/yaKhzrgFKMK/0TcCF/DAr6dB4uHxjWoct0bzS+fQYbwEOAL4GMLP38UcCnXNZVCF96nwz65M23bm2Dh0AzDWzcRnV/w7oBvQF1gcuSq1SQVNqNK90iZl9mTFvdR7rOecamKTSpwI/AQ6UNJUwaM0ASQ+Z2ax4uLycMLhNv1i+1vNKT5PUD7B4hegcYFIe6znnGqBGym/Kxcx+Z2YdzKwjMBh4xcyOiecNieM3HMzaAbVrN680cBrhcHpzYA7wUpznnHPlqOZHznlYUltC53Q8cGqcX7t5pc1sLiFiO+dcpZKOi2b2KvBqfD0gS5lE80rnM4L3XVRwstLMMq8gOedcUYzgnc+h9Etpr5sBgyh/+ds554AGkD41xcweTX8v6UHCTZLOOVeeGk6PMVMnYIukG+KcKw6q8NbB+iWfc4wLWXuOsQRYAFxck41yztVPxZIlMGdgjPcK9SDkeQFYE6/+OOdchYo+MJqZSXrKzHrXVoOcc/VXsVx8yefJl7fTxzZzzrms8nxOuq4P8p0r50upma0CdgNOljQF+JbwR8HMzIOlc+4H6nrO6HzkOpR+G+hFeB7ROecq1RAuvgjAzKbUUlucc/WeaFTkPca2ks7LttDM/loD7XHO1WMh50uhW/Hj5QqMjYCWVDzQo3PO/VADePJllpldWWstcc4VhWK/+FL/9845V6uK5VA6132Me9RaK5xzRaNRifKa8lFBXulOkt6KOaIfldQkzm8a30+Oyzum1fG7OP9TSfvks92sgTGfUW6dcy6dCEElnylPqbzSKX8GbjSzLsBC4KQ4/yRgoZl1Bm6M5ZC0DWGg7W0JaVNvl9Soso0mmxrbOdew1WBe6Th2wwDg8VjkAdbeZ31QfE9cvkcsfxAwzMyWm9kXhNQHqQRaWXlgdM4lSnlOVD2v9AbAovhEHpTPEV2WPzouXxzLVyuvdHXGY3TOuQql0qfmab6Z9amwnrS80pL6p1WfySpZVq280h4YnXOJSug+xlRe6f0IKVVaE3qQ66WN45CeIzqVP3q6pFJgXcLYsTWWV9o55/KU3/nFys4xZskrfTTwH+DQWGwI8HR8PSK+Jy5/JY4dOwIYHK9adwK6EMaByMl7jM65xKSuStegi4Bhkq4G3gPuifPvAR6UNJnQUxwMYGYTJA0HPgZWAWeY2erKNuKB0TmXqHyuOFdFRl7pz6ngqrKZfQ8clmX9ocDQqmzTA6NzLjkq/kcCnXOuSmrhULpWeGB0ziUq6UPpQvDA6JxLVP0Pix4YnXMJK4IOowdG51xyBEWf2sA556pIqAgOpj0wOucSVQQdRg+MzrnkhNt16n9k9MDonEuOoKQIbmT0wOicS1QxnGMsgtheN82YPo2D99uTXXtvz259e/D3228pW3bXHbey847bslvfHlxx6cUAPP7oI/TftXfZtFHrJnz4wXgADh+0P/136cVufXtw/tmns3p1pc/Auzytu05THvnDgYy/5wTeu/sEduq+admycw7tw3cvns8GrZsDsPVm6/PqTUex6NlzOOfQ8sMI7tWnI+/fcyIf3XcS5x9R6QDRRSuMx5jfVJd5j7GGNCot5Yo/XUePnr1YumQJe+y+E/0H7Mm8uXN54blneG3MuzRt2pR58+YCcOgRR3HoEUcB8PGEDzlu8CFsv0NPAO554J+0at0aM+OEY45gxFOPM+jQIwq2b8Xk+tMH8OLYLzjqqhE0Li2hRdPGAHRo24oBvbbgqznflJVduOR7fnv7K/xi187l6igpETeduSf7X/wYM+Yv4Y3/O4ZnR09h4ldf1+q+1BXeY3RZbbLJpvTo2QuAlq1asXXXbsyaOZP77v47vznvQpo2bQpA27Yb/WDdJx97tFzga9W6NQCrVq1i5YoVxXHZrw5o1aIJu23fgftf+BCAlavWsPjb5QBcd+rPueTu/xKG9AvmLVrGuEmzWbl6Tbl6+nbdhCkzFzJ19mJWrlrDY69N5IBdt6q9HaljpPymuswDYy346supfPjBeHr36ceUyZMY87832Ofnu3LgwAG8N27sD8o//eRj/PKw8j3Cww7ej+5btqNlq1YcePAhtdX0otZpk3WZv2gZd54/kNG3H8vt5+5Ni2aN2X/nrZg5fwkffj4vr3rabdiK6fOWlL2fMW8p7TdoVVPNrtNSN3jnM9VlHhhr2NKlSznhmMO5+tobaNW6NatXrWbRooW88MqbXH71tfxqyFHleiXjxr5F8+bN6b7NduXqeexfz/PRZ9NYvnw5r7/2n9rejaJU2qiEnl025q5nx7PL6Q+y7PuVXHrsrlx01M5c+cCbeddTYVIRqzStSJFS3v9y1iI1k/S2pPclTZB0RZx/v6QvJI2PU884X5JuifmjP5DUK62uITEP9WeShmTbZjoPjDVo5cqVnHDM4Rx6+JEccNAgADZt354DDhyEJHr16UdJSQlfz59fts5TTwxn0KGDK6yvWbNmDNzvAP793IhaaX+xmzF/CTPmLWHsxNkAPPX6JHp23ogtNlmXt+8YwsR/nEz7tq0YffuxbNymRc56OrRd20Ns37YlMxcsrfH210l5Hkbn0WFcDgwwsx5AT2CgpJ3jsgvMrGecxsd5+xLSFnQBTgH+BiBpfeAyYCfCALeXSWpT2cZrNTBK6ijpE0l3xb8CL0pqLqmnpDEx0j+VT8PrOjPjnDNOZuuu3TjtrHPL5u93wIFlPb4pn01ixYoVbLDhhgCsWbOGEU89waBDDy8rv3TpUmbPngWEc4wvvfgCXbbuWot7UrzmLFzG9HlL6NIhfN3677gF4yfPZYvDb6fbcXfR7bi7mDFvCbuc/iBzFi7LWs87n86mc/s2bLHJujQuLeGwn3XjudFTams36pwqpE/NyoLUX5fGccrVDT8I+EdcbwwhadamwD7AKDNbYGYLgVHAwMr2oRBXpbsAR5rZyTEXwyGE3LFnmdlrkq4kRPhz0leKOWdPAeiw2ea13OSqe2v0mwz/58Nss+129N+91daEAAAMK0lEQVS1NwCXXHY1Rx17Amef/it279eTxk0ac+vf7y0bv270m6/Trl17OnbasqyeZcu+5dgjBrFi+XJWr17Dbj/rz/En/bog+1SMzrvtZe67eH+alDZi6uxFnHL9C1nLbtymBW/eeiytWjRhjRlnDurNjiffx5JlKzj31pd55k+H0KikhAdGfsgnXzbUK9JVGkRiQ0nvpL2/08zuLKtLagSMAzoDt5nZW5JOA4ZK+iPwMnCxmS0ne/7oauWVVm2eC5HUkRC9u8T3FxFSI55kZpvHeVsBj5lZr2z19OzV217671s132BXbZv98qZCN8Hl4ftRF4zLltu5Orpvv6Pd96/8zoHv0rlNXtuWtB7wFHAW8DUwG2gC3AlMMbMrJT0HXGNmb8R1XiZ0uAYATc3s6jj/D8AyM7sh1zYLcY5xedrr1cB6BWiDc66GJHHxJZ2ZLSIkwxpoZrPi4fJy4D7WJsbKlj+63uaVXgwslLR7fH8s8FoB2+Oc+xGSuPgiqW3sKSKpObAnMDGeN0Th/NPBwEdxlRHAcfHq9M7AYjObBYwE9pbUJl672DvOy6muPPkyBLhDUgvgc+CEArfHOVdNCd2huCnwQDzPWAIMN7NnJb0iqW3czHjg1Fj+eWA/YDKwjBhDzGyBpKuA1A3DV5rZgso2XquB0cymAtulvb8+bfHOP1jBOVeviGSSYZnZB8COFcwfkKW8AWdkWXYvcG9Vtl9XeozOuWJQDx73y4cHRudcooogLnpgdM4lrAgiowdG51yCPBmWc86Vkxqotr7zwOicS5YHRuecK88PpZ1zLoPfruOcc+n8PkbnnPshP5R2zrk04ZHAQrfix/PA6JxLVBHERQ+MzrmEFUFk9MDonEtUSREcS3tgdM4lqv6HxboxgrdzrpgkkCYwR17pTpLeijmiH5XUJM5vGt9Pjss7ptX1uzj/U0n75LMLHhidc4kJMS+RnC/Z8kr/GbgxJtRbCJwUy58ELDSzzsCNsRyStgEGA9sS0qbeHkcFz8kDo3MuOQqDSOQz5ZIjr/QA4PE4/wFC3hcIeaUfiK8fB/aIeWEOAoaZ2XIz+4KQ+iCVQCsrD4zOuWTlfyi9oaR30qZTylUjNZI0HpgLjAKmAIvMbFUskp4juix/dFy+GNiAauaV9osvzrkEVWk8xvm58kqb2WqgZ1pe6e4VFSvbcMXLss3PyXuMzrlEJZE+NV1aXumdgfUkpTp06Tmiy/JHx+XrAguox3mlnXNFIt+j6MriYpa80p8A/wEOjcWGAE/H1yPie+LyV2LmwBHA4HjVuhPQBXi7sv3wQ2nnXKKSSJ9K9rzSHwPDJF0NvAfcE8vfAzwoaTKhpzgYwMwmSBoOfAysAs6Ih+g5eWB0ziUqibiYI6/051RwVdnMvgcOy1LXUGBoVbbvgdE5l6hiePLFA6NzLjk+UK1zzlWk/kdGD4zOucR4+lTnnKuAH0o751wGz/ninHOZ6n9c9MDonEuO8hg5pz7wwOicS5QfSjvnXKb6Hxc9MDrnklUEcdEDo3MuWX67jnPOpREqivSpPh6jc85l8B6jcy5RRdBh9B6jcy5ZSaRPlbSZpP9I+iTmlT47zr9c0gxJ4+O0X9o6FeaPljQwzpss6eJ89sF7jM65xCR4g/cq4Ldm9q6kVsA4SaPishvN7Pry2y2XP7od8JKkrePi24C9CPlfxkoaYWYf59q4B0bnXLKSGcF7FjArvl4i6RNypz0tyx8NfBFTHKRG+p4cR/5G0rBYNmdg9ENp51yiqnAonTOvdFl9UkdCmoO34qwzJX0g6V5JbeK8bPmjq5VX2gOjcy5RVUifOt/M+qRNd/6wLrUEngDOMbNvgL8BWwE9CT3KG1JFK2hKtfNK+6G0cy5RSV2UltSYEBQfNrMnAcxsTtryu4Bn49tc+aM9r7RzrrAk5TVVUocIKVE/MbO/ps3fNK3YIOCj+Dpb/uixQBdJnSQ1IVygGVHZPniP0TmXGJHYfYw/AY4FPpQ0Ps77PXCkpJ6Ew+GpwK8hd/5oSWcCI4FGwL1mNqHS/TCr9HC7zpE0D/iy0O1I2IbA/EI3wuVUjJ/RFmbWNqnKJL1A+DnlY76ZDUxq20mql4GxGEl6x8z6FLodLjv/jBoOP8fonHMZPDA651wGD4x1xw/u4XJ1jn9GDYSfY3TOuQzeY3TOuQweGJ1zLoMHRuecy+CB0TnnMnhgrGMkbSqpRaHb4SonyX9/ipR/sHWIpAMJwypVOl6cq32SjpZ0qaSzJW1uZms8OBYn/1DrCEm7A1cAfzSzzyQ1k9Q6LiuC9EL1m6QzgLOAJcAWwBOSOpvZmsK2zNUEH12nwCTJws2k3YHXgNWSTgcGAt9LusDMim3AjHoj7fPZHviNmb0d518E/EHSqWb2XUEb6RLnPcbCaxX/Hws0Bx4jDKl0D/AFsF6B2uWCLnHA1A5A/7T5/wZWeFAsTt5jLCBJ+xPGl/scGAdcDJSY2deSdgSuBR4pZBsbsjiO3znAU8D7wG8kzTezewk9yK0krWtmiwvZTpc8D4wFIqkvcB1wMKF32BF4ISzSbsB9wLlm9n7BGtmAxQthOwD7AHsDrYGXgKvjH62fA0d4UCxO/qx0AUjanBAQ5xIG3L0ZOMzMvpTUnjDQZ2Mze6eAzWyw4mcwGnjJzE6U1BQ4hJA7pA1hMInFZvZ1AZvpapCfY6xlkjYGziSMBH0K4ZdsUAyKhwKnE/JceFAsEDObQTiEHihpcMxVPAyYB6wBFnhQLG5+KF375gNbA1sCnwIvAq0ltQP+AFxqZisK2D4HmNmTkpYD10jCzIZJuh9Yx8yWFLh5rob5oXQtiYGvpZlNiofS5wOTgA0I56uWAneZ2dNpt4i4ApO0L6FXf66ZPV7o9rja4YGxFkhaB7ga6EE4JBtNOGR+0Mz+J6kV4ZziAg+KdY+kvYApZvZ5odviaocHxloiqRmwDXAR8AHhHNZU4JdmNq2ATXPOZfBzjLXEzL4H3pV0CtCUcOGrJ+HG4WneU3Su7vAeYwFJuoSQ1/eUQrfFObeW365TAGmDQkwBtpDUvJDtcc6V54GxAMzMYnD8FvitP2/rXN3ih9LOOZfBe4zOOZfBA6NzzmXwwOiccxk8MBYJSasljZf0kaTHfkxCLUn9JT0bXx8o6eIcZdeLI45XdRuXSzo/3/kZZe6PA27ku62Okj6qahtdw+WBsXh8Z2Y9zWw7YAVwavpCBVX+vM1shJldm6PIeoTHG50rGh4Yi9PrQOfYU/pE0u3Au8BmkvaWNFrSu7Fn2RJA0kBJEyW9AfwyVZGk4yXdGl9vLOkpSe/HaVfCKONbxd7qX2K5CySNlfSBpCvS6rpE0qeSXgK6VrYTkk6O9bwv6YmMXvCekl6XNEnSAbF8I0l/Sdv2r3/sD9I1TB4Yi4ykUmBf4MM4qyvwDzPbkXDf5KXAnmbWC3gHOC8+x30X8Atgd2CTLNXfArxmZj2AXsAEQjqGKbG3eoGkvYEuQD/CI4+9Jf1UUm9gMLAjIfD2zWN3njSzvnF7nwAnpS3rCPwM2B+4I+7DSYQBZPvG+k+W1CmP7ThXjj8rXTyaSxofX79OSJfQDvjSzMbE+TsTBrJ4Mz5804Qw0k834Asz+wxA0kOEQXQzDQCOAzCz1cBiSW0yyuwdp/fi+5aEQNkKeMrMlsVtjMhjn7aTdDXhcL0lMDJt2fCYuvSzmDOnW9zuDmnnH9eN256Ux7acK+OBsXh8Z2Y902fE4Pdt+ixglJkdmVGuJyEzYRIEXGNmf8/YxjnV2Mb9wMFm9r6k4ymfpS+zLovbPsvM0gMokjpWcbuugfND6YZlDPATSZ0BJLWQtDUwEegkaatY7sgs678MnBbXbSSpNSEBfau0MiOBE9POXbaXtBHwX2CQpOZx/Mlf5NHeVsAshfSlR2csO0xSSWxzajT0kcBpsTySto5jYTpXJd5jbEDMbF7sef1TIcEThFQKk+JwaM9Jmg+8AWxXQRVnA3dKOglYDZxmZqMlvRlvh/l3PM/YHRgde6xLgWPM7F1JjwLjCQnAXs+jyX8A3orlP6R8AP4UeA3YGDjVzL6XdDfh3OO78Vn0eYSkY85ViT8r7ZxzGfxQ2jnnMnhgdM65DB4YnXMugwdG55zL4IHROecyeGB0zrkMHhidcy7D/wPCNWuxDn6WdAAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "\n",
    "# Evaluate the confusion matrix\n",
    "int_to_class = {0: \"yes\", 1: \"no\"}\n",
    "cnf_matrix = confusion_matrix([int_to_class[c] for c in y_test], [int_to_class[c] for c in pred])\n",
    "\n",
    "# Plot the confusion matrix\n",
    "plt.figure()\n",
    "plot_confusion_matrix(cnf_matrix, classes=class_names,\n",
    "                      title='Confusion matrix, with normalization')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# random forest\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.25, random_state=42)\n",
    "x_plot=[]\n",
    "y_plot=[]\n",
    "for step in range(1, 50):\n",
    "    rf_model=RandomForestClassifier(n_estimators=step)\n",
    "    rf_model.fit(X_train, y_train)\n",
    "    pred=rf_model.predict(X_test)\n",
    "    acc=accuracy_score(y_test, pred)\n",
    "    x_plot.append(step)\n",
    "    y_plot.append(acc)\n",
    "\n",
    "draw_acc(x_plot=x_plot,y_plot=y_plot)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### n_estimators=49\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "final result : 0.933979 \n"
     ]
    }
   ],
   "source": [
    "model=[]\n",
    "res=[]\n",
    "for step in range(0, 15):\n",
    "    X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.25, random_state=step) #random select\n",
    "    rf_model=RandomForestClassifier(n_estimators=49)\n",
    "    rf_model.fit(X_train, y_train)\n",
    "    pred=rf_model.predict(X_test)\n",
    "    acc=accuracy_score(y_test, pred)\n",
    "    model.append(rf_model)\n",
    "    res.append(acc)\n",
    "avg_acc=sum(res)/15.\n",
    "print ('final result : %f ' %avg_acc)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Confusion matrix, without normalization\n",
      "[[8728  463]\n",
      " [ 668 8415]]\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "int_to_class = {0: \"yes\", 1: \"no\"}\n",
    "cnf_matrix = confusion_matrix([int_to_class[c] for c in y_test], [int_to_class[c] for c in pred])\n",
    "\n",
    "# Plot the confusion matrix\n",
    "plt.figure()\n",
    "plot_confusion_matrix(cnf_matrix, classes=class_names,\n",
    "                      title='Confusion matrix, with normalization')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# KNN"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.25, random_state=21)\n",
    "x_plot=[]\n",
    "y_plot=[]\n",
    "for step in range(1, 30):\n",
    "    knn_model = KNeighborsClassifier(n_neighbors=step)\n",
    "    knn_model.fit(X_train, y_train)\n",
    "    pred=knn_model.predict(X_test)\n",
    "    acc=accuracy_score(y_test, pred)\n",
    "    x_plot.append(step)\n",
    "    y_plot.append(acc)\n",
    "draw_acc(x_plot=x_plot,y_plot=y_plot)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "final result : 0.857025 \n"
     ]
    }
   ],
   "source": [
    "model=[]\n",
    "res=[]\n",
    "for i in range(0, 15):\n",
    "    X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.25, random_state=i)\n",
    "    knn_model = KNeighborsClassifier(n_neighbors=1)\n",
    "    knn_model.fit(X_train, y_train)\n",
    "    pred=knn_model.predict(X_test)\n",
    "    acc=accuracy_score(y_test, pred)\n",
    "    model.append(knn_model)\n",
    "    res.append(acc)\n",
    "    \n",
    "\n",
    "avg_acc=sum(res)/15.\n",
    "print ('final result : %f ' %avg_acc)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Confusion matrix, without normalization\n",
      "[[7441 1750]\n",
      " [ 841 8242]]\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Evaluate the confusion matrix\n",
    "cnf_matrix = confusion_matrix([int_to_class[c] for c in y_test], [int_to_class[c] for c in pred])\n",
    "\n",
    "# Plot the confusion matrix\n",
    "plt.figure()\n",
    "plot_confusion_matrix(cnf_matrix, classes=class_names,\n",
    "                      title='Confusion matrix, without normalization')\n",
    "plt.show()"
   ]
  }
 ],
 "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": "3.7.1"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}