LinearModels_Baseline.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "provenance": [],
      "collapsed_sections": [],
      "authorship_tag": "ABX9TyNjddCHj6BtF3CVyEFpLpKc",
      "include_colab_link": true
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "language_info": {
      "name": "python"
    }
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "view-in-github",
        "colab_type": "text"
      },
      "source": [
        "<a href=\"https://colab.research.google.com/github/jacobdwatters/NIOSH-Project/blob/main/LinearModels_Baseline.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "## Imports and setup"
      ],
      "metadata": {
        "id": "kjCUrngFsVEI"
      }
    },
    {
      "cell_type": "code",
      "execution_count": 335,
      "metadata": {
        "id": "ekB0Ix6IYvoi"
      },
      "outputs": [],
      "source": [
        "import numpy as np\n",
        "import pandas as pd\n",
        "import seaborn as sns\n",
        "import matplotlib.pyplot as plt\n",
        "\n",
        "from sklearn.linear_model import LinearRegression\n",
        "from sklearn.preprocessing import StandardScaler\n",
        "from sklearn.preprocessing import OneHotEncoder\n",
        "from sklearn.model_selection import train_test_split\n",
        "from sklearn.linear_model import RANSACRegressor\n",
        "from sklearn.metrics import mean_absolute_error\n",
        "from sklearn.metrics import r2_score\n",
        "\n",
        "import scipy as sp\n",
        "from scipy import stats"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "from google.colab import drive\n",
        "drive.mount('/content/gdrive')"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 0
        },
        "id": "D5xBDHi_aZhH",
        "outputId": "5201df95-4d9b-4630-9bc3-8efbd11ce1db"
      },
      "execution_count": 336,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Drive already mounted at /content/gdrive; to attempt to forcibly remount, call drive.mount(\"/content/gdrive\", force_remount=True).\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "## Load and proceses data"
      ],
      "metadata": {
        "id": "GuicPgXfaAH9"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "#### Load data from csv"
      ],
      "metadata": {
        "id": "p6EWQRv3dLZn"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "path_after_2010 = '/content/gdrive/My Drive/NIOSH Project/data/violations_processed_after_2010.csv'\n",
        "violation_data = pd.read_csv(path_after_2010)"
      ],
      "metadata": {
        "id": "zzkxoejnZHTu"
      },
      "execution_count": 337,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "#### Split into features and targets"
      ],
      "metadata": {
        "id": "LMpnBopCdOSD"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "FEATURES = ['MINE_TYPE', 'COAL_METAL_IND', 'SIG_SUB', 'LIKELIHOOD', \n",
        "            'INJ_ILLNESS', 'NO_AFFECTED', 'NEGLIGENCE', 'VIOLATOR_VIOLATION_CNT',\n",
        "            'VIOLATOR_INSPECTION_DAY_CNT']\n",
        "TARGETS = ['PROPOSED_PENALTY']\n",
        "\n",
        "X = violation_data[FEATURES]\n",
        "y = violation_data[TARGETS]"
      ],
      "metadata": {
        "id": "kFB2mKZKahOt"
      },
      "execution_count": 338,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "#### Encode and scale features"
      ],
      "metadata": {
        "id": "-heZ96xddR9A"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# Define which columns should be encoded vs scaled\n",
        "columns_to_encode = ['MINE_TYPE', 'COAL_METAL_IND', 'LIKELIHOOD', 'INJ_ILLNESS', 'SIG_SUB', 'NEGLIGENCE']\n",
        "columns_to_scale  = ['VIOLATOR_VIOLATION_CNT', 'NO_AFFECTED', 'VIOLATOR_INSPECTION_DAY_CNT']\n",
        "\n",
        "# Instantiate encoder/scaler\n",
        "scaler = StandardScaler()\n",
        "ohe = OneHotEncoder(sparse=False)\n",
        "\n",
        "# Scale and Encode Separate Columns\n",
        "scaled_columns  = scaler.fit_transform(X[columns_to_scale])\n",
        "encoded_columns = ohe.fit_transform(X[columns_to_encode])\n",
        "\n",
        "# Concatenate (Column-Bind) Processed Columns Back Together\n",
        "X_pre = np.concatenate([scaled_columns, encoded_columns], axis=1)\n",
        "np.nan_to_num(X_pre, copy=False)\n",
        "\n",
        "print('Features shape:', X_pre.shape)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 0
        },
        "id": "s7I23o80bfIe",
        "outputId": "8aa49943-b01b-457c-b534-b46ad802069d"
      },
      "execution_count": 339,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Features shape: (1429135, 24)\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "#### Split data into training and testing datasets"
      ],
      "metadata": {
        "id": "n6z0SgrWdq_b"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "X_train, X_test, y_train, y_test = train_test_split(X_pre, y, test_size = 0.25, random_state = 0)\n",
        "\n",
        "print('X_train shape:', X_train.shape)\n",
        "print('X_test shape:', X_test.shape)\n",
        "print('y_train shape:', y_train.shape)\n",
        "print('y_train shape:', y_test.shape)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 0
        },
        "id": "7EQnPj02dW4I",
        "outputId": "37e4e765-8e71-44c6-f1a9-13f885c5e2a5"
      },
      "execution_count": 340,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "X_train shape: (1071851, 24)\n",
            "X_test shape: (357284, 24)\n",
            "y_train shape: (1071851, 1)\n",
            "y_train shape: (357284, 1)\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "## Define and Fit Simple Linear Regression Model"
      ],
      "metadata": {
        "id": "AhYAmoLreB3G"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "lin_reg = LinearRegression().fit(X_train, y_train)"
      ],
      "metadata": {
        "id": "iplwrboHeE_v"
      },
      "execution_count": 341,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "lin_reg_y_pred_train = lin_reg.predict(X_train)\n",
        "lin_reg_y_pred_test = lin_reg.predict(X_test)\n",
        "\n",
        "lin_reg_r_squared_train = r2_score(y_train, lin_reg_y_pred_train)\n",
        "lin_reg_r_squared_test = r2_score(y_test, lin_reg_y_pred_test)\n",
        "\n",
        "lin_reg_mae_train = mean_absolute_error(y_train, lin_reg_y_pred_train)\n",
        "lin_reg_mae_test = mean_absolute_error(y_test, lin_reg_y_pred_test)\n",
        "\n",
        "print('Simple Linear Regression Scores:')\n",
        "print('Training R^2 = %.3f' % lin_reg_r_squared_train)\n",
        "print('Training MAE = %.3f' % lin_reg_mae_train)\n",
        "\n",
        "print('\\nTesting R^2 = %.3f' % lin_reg_r_squared_test)\n",
        "print('Testing MAE = %.3f' % lin_reg_mae_test)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 0
        },
        "id": "ND7kRoBjh1tF",
        "outputId": "11177bd2-e013-47b4-85a1-6bbd7a097f08"
      },
      "execution_count": 342,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Simple Linear Regression Scores:\n",
            "Training R^2 = 0.276\n",
            "Training MAE = 841.728\n",
            "\n",
            "Testing R^2 = 0.260\n",
            "Testing MAE = 837.038\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "## Define and Fit RANSAC Regressor"
      ],
      "metadata": {
        "id": "to1YjmuRfgRd"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "ransac_reg = RANSACRegressor().fit(X_train, y_train)"
      ],
      "metadata": {
        "id": "xeGAQ90Uffre"
      },
      "execution_count": 343,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "ransac_reg_y_pred_train = ransac_reg.predict(X_train)\n",
        "ransac_reg_y_pred_test = ransac_reg.predict(X_test)\n",
        "\n",
        "ransac_reg_r_squared_train = r2_score(y_train, ransac_reg_y_pred_train)\n",
        "ransac_reg_r_squared_test = r2_score(y_test, ransac_reg_y_pred_test)\n",
        "\n",
        "ransac_reg_mae_train = mean_absolute_error(y_train, ransac_reg_y_pred_train)\n",
        "ransac_reg_mae_test = mean_absolute_error(y_test, ransac_reg_y_pred_test)\n",
        "\n",
        "print('RANSAC Regression Scores:')\n",
        "print('Training R^2 = %.3f' % ransac_reg_r_squared_train)\n",
        "print('Training MAE = %.3f' % ransac_reg_mae_train)\n",
        "\n",
        "print('\\nTesting R^2 = %.3f' % ransac_reg_r_squared_test)\n",
        "print('Testing MAE = %.3f' % ransac_reg_mae_test)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 0
        },
        "id": "HbO5Sbuth5_8",
        "outputId": "97b5c555-e0a0-47b7-f892-62c34f13c5ed"
      },
      "execution_count": 344,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "RANSAC Regression Scores:\n",
            "Training R^2 = -0.011\n",
            "Training MAE = 575.063\n",
            "\n",
            "Testing R^2 = -0.012\n",
            "Testing MAE = 565.638\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "# Removing Outliers in Target Data\n",
        "Most of the data in the PROPOSED_PENALTY column seems to be less than 500. So, let's try removing large values from the dataset."
      ],
      "metadata": {
        "id": "WijkEZAolRFr"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "MAX_VALUE = 1000\n",
        "violations_cliped = violation_data[violation_data['PROPOSED_PENALTY'] <= MAX_VALUE]\n",
        "\n",
        "print('Total samples = %d' % len(violation_data))\n",
        "print('Total samples with PROPOSED_PENALTY <= %d: %d' % (MAX_VALUE, len(violations_cliped)))"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 0
        },
        "id": "yjNm-iuMkRPd",
        "outputId": "c983dbb9-0789-4cd3-910b-6b53008cb3f6"
      },
      "execution_count": 345,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Total samples = 1429135\n",
            "Total samples with PROPOSED_PENALTY <= 1000: 1259626\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "hist = violations_cliped.hist(figsize=(10,10), bins=40)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 607
        },
        "id": "NFoAuX_ymW8t",
        "outputId": "7459ac48-c7b2-4677-cddf-0d9f26f1b549"
      },
      "execution_count": 346,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 720x720 with 9 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "## Prepare Inlier Data"
      ],
      "metadata": {
        "id": "-VnILmiXnMug"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "X_inliers = violations_cliped[FEATURES]\n",
        "y_inliers = violations_cliped[TARGETS]\n",
        "\n",
        "# Instantiate encoder/scaler\n",
        "scaler = StandardScaler()\n",
        "ohe = OneHotEncoder(sparse=False)\n",
        "\n",
        "# Scale and Encode Separate Columns\n",
        "scaled_columns_inliers  = scaler.fit_transform(X_inliers[columns_to_scale])\n",
        "encoded_columns_inliers = ohe.fit_transform(X_inliers[columns_to_encode])\n",
        "\n",
        "# Concatenate (Column-Bind) Processed Columns Back Together\n",
        "X_pre_inliers = np.concatenate([scaled_columns_inliers, encoded_columns_inliers], axis=1)\n",
        "np.nan_to_num(X_pre_inliers, copy=False)\n",
        "\n",
        "X_train_inliers, X_test_inliers, y_train_inliers, y_test_inliers = train_test_split(X_pre_inliers, y_inliers, test_size = 0.25, random_state = 0)"
      ],
      "metadata": {
        "id": "5omO5YwenL0r"
      },
      "execution_count": 347,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "## Inlier Linear Regression"
      ],
      "metadata": {
        "id": "ATp2SEDXn5x3"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "lin_reg_inliers = LinearRegression().fit(X_train_inliers, y_train_inliers)"
      ],
      "metadata": {
        "id": "E6RnxsACn44s"
      },
      "execution_count": 348,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "lin_reg_inliers_y_pred_train = lin_reg_inliers.predict(X_train_inliers)\n",
        "lin_reg_inliers_y_pred_test = lin_reg_inliers.predict(X_test_inliers)\n",
        "\n",
        "lin_reg_inliers_r_squared_train = r2_score(y_train_inliers, lin_reg_inliers_y_pred_train)\n",
        "lin_reg_inliers_r_squared_test = r2_score(y_test_inliers, lin_reg_inliers_y_pred_test)\n",
        "\n",
        "lin_reg_inliers_mae_train = mean_absolute_error(y_train_inliers, lin_reg_inliers_y_pred_train)\n",
        "lin_reg_inliers_mae_test = mean_absolute_error(y_test_inliers, lin_reg_inliers_y_pred_test)\n",
        "\n",
        "print('Inlier Simple Linear Regression Scores:')\n",
        "print('Training R^2 = %.3f' % lin_reg_inliers_r_squared_train)\n",
        "print('Training MAE = %.3f' % lin_reg_inliers_mae_train)\n",
        "\n",
        "print('\\nTesting R^2 = %.3f' % lin_reg_inliers_r_squared_test)\n",
        "print('Testing MAE = %.3f' % lin_reg_inliers_mae_test)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 0
        },
        "id": "RZ5grPRAn_ua",
        "outputId": "99075b4a-0ae2-4f8c-d70d-5983c8d4035c"
      },
      "execution_count": 349,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Inlier Simple Linear Regression Scores:\n",
            "Training R^2 = 0.388\n",
            "Training MAE = 105.540\n",
            "\n",
            "Testing R^2 = 0.388\n",
            "Testing MAE = 105.501\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "## Inlier RANSAC Regression"
      ],
      "metadata": {
        "id": "SKfYNjawoSZE"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "ransac_reg_inliers = RANSACRegressor(random_state=35).fit(X_train_inliers, y_train_inliers)"
      ],
      "metadata": {
        "id": "SamR2uc9oYGi"
      },
      "execution_count": 350,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "ransac_reg_inliers_y_pred_train = ransac_reg_inliers.predict(X_train_inliers)\n",
        "ransac_reg_inliers_y_pred_test = ransac_reg_inliers.predict(X_test_inliers)\n",
        "\n",
        "ransac_reg_inliers_r_squared_train = r2_score(y_train_inliers, ransac_reg_inliers_y_pred_train)\n",
        "ransac_reg_inliers_r_squared_test = r2_score(y_test_inliers, ransac_reg_inliers_y_pred_test)\n",
        "\n",
        "ransac_reg_inliers_mae_train = mean_absolute_error(y_train_inliers, ransac_reg_inliers_y_pred_train)\n",
        "ransac_reg_inliers_mae_test = mean_absolute_error(y_test_inliers, ransac_reg_inliers_y_pred_test)\n",
        "\n",
        "print('Inlier RANSAC Regression Scores:')\n",
        "print('Training R^2 = %.3f' % ransac_reg_inliers_r_squared_train)\n",
        "print('Training MAE = %.3f' % ransac_reg_inliers_mae_train)\n",
        "\n",
        "print('\\nTesting R^2 = %.3f' % ransac_reg_inliers_r_squared_test)\n",
        "print('Testing MAE = %.3f' % ransac_reg_inliers_mae_test)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 0
        },
        "id": "Y7GGKXGmoZdM",
        "outputId": "42e4ff95-3199-4966-95fe-fba1272331a6"
      },
      "execution_count": 351,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Inlier RANSAC Regression Scores:\n",
            "Training R^2 = 0.090\n",
            "Training MAE = 112.170\n",
            "\n",
            "Testing R^2 = 0.102\n",
            "Testing MAE = 111.643\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "# Focusing on Outliers\n",
        "Now lets see if we can fit a model to the larger values."
      ],
      "metadata": {
        "id": "xaf0vu3gqhx1"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "violations_outliers = violation_data[violation_data['PROPOSED_PENALTY'] > MAX_VALUE]\n",
        "\n",
        "print('Total samples = %d' % len(violation_data))\n",
        "print('Total samples with PROPOSED_PENALTY > %d: %d' % (MAX_VALUE, len(violations_outliers)))"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 0
        },
        "id": "5CqDGMK4quNX",
        "outputId": "c8607ef5-ce04-4f35-f8fa-a2611f807ca7"
      },
      "execution_count": 352,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Total samples = 1429135\n",
            "Total samples with PROPOSED_PENALTY > 1000: 169509\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "hist = violations_outliers.hist(figsize=(10,10), bins=40)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 607
        },
        "id": "MerpGeNFq-YG",
        "outputId": "686bb8f5-6aa5-4001-810e-2c7985c1c34b"
      },
      "execution_count": 353,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 720x720 with 9 Axes>"
            ],
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAn8AAAJOCAYAAADGaGHzAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4yLjIsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+WH4yJAAAgAElEQVR4nOzde7gcVZ3v//dHLoLIJRDdxgQNP4nOABkQMhBHZyaChIBo8IgIMiY4AeaMoMwxjgbHI8hlBB1EUEQRIgHBEFEkRwIxAnuUc0y4CBoCIhGiJAYiSQhsrga+vz/W6qR2p3vv3pe+7N2f1/P0s7tWrar6VnVV99pV66KIwMzMzMzaw6uaHYCZmZmZNY4Lf2ZmZmZtxIU/MzMzszbiwp+ZmZlZG3Hhz8zMzKyNuPBnZmZm1kZc+LNBJWmspJC0dbNjMTMzsy258NckuYC0Z1namZK+16yYmkHSrpJukPSspD9I+kizYzIzMxvOXPizZrsEeAnoAI4HLpW0d3NDsqFK0gpJL0kaWZZ+b/6Ha6ykKyWdk9NLd6oXlOX/nqQz8/tJkl6R1FX2ekcvsXRKOrHKOlZKmifpbwf1ANiwlc/tNZJ2KKSdKKkzv5ekf5f0sKTnJf1R0pckvboP2zgzXw8HlaWfIOnlsvP/G3nelfmaK877cCHm5ystl+ePknSFpNWSnpH0W0lflPTmsmUi3yAoTf99le3+Oq+3dF2X0p+Q9BNJhw7oQxhGXPhrUfnHYqWkmfmCXy3pY4X5V0q6RNJN+aJZIukthfkXSXpM0tOS7pH094V5Z0r6Qf6Be0bSUklvlXR63tZjkiYX8u9cuEBXSTpH0lZ53laS/kvSk5IeAd7bh33cAfgg8L8joisi7gDmAx8d0MGzdvcocFxpQtJ44DW9LHOQpL/rYf6fIuK1Za9f9jGuP0XEa4EdgYnAb4FfSDqkj+ux9rUVcFqVeRcDJwPTSOfY4cAhwLxaVixJedl1+W+5X5ad/6cW5n25bN51hXnvq7ScpF2BXwLbA++IiB2BQ4FdgJ2Ly+T17FtI+0WV7e5bFvMuefl9gUXADZJOqOV4DHcu/LW2NwA7A6OBGcAlkkYU5h8LfBEYASwHzi3MuwvYD9gVuBb4gaTtCvPfB1ydl70XWEg6H0YDZwHfLuS9EtgI7Am8HZgMnJjnnQQcmdMnAEcXd0DSLEk/qbJ/bwU2RsTvCmm/Bnznzwbiarr/eE0HruplmS/T/fqpm0hWRsQXgMuB8xuxXRsWvgJ8WtIuxURJ44CPA8dHxC8jYmNELCP9cz1F0sE1rPvvgVHAJ4FjJW07yLGX+xTwDPBPEbECICIei4jTIuI3g7mhiHg8Ii4CzgTOl9T2ZZ+2PwAt7i/AWRHxl4hYAHQBbyvMvyEi7oyIjcA1pMIeABHxvYhYm78ELgBeXbbsLyJiYV72B8DrgPMi4i/AXGCspF0kdQBHAP8WEc9GxBrgQlLBE+AY4Gv5ol0HfKm4AxFxXkQcWWX/Xgs8XZa2gfRfq1l/LQZ2kvTX+Q71sUBvdWm/CbxV0nvqHl13PwL2Lz7KM+vB3UAn8Omy9EOAlRFxZzExIh4jXQ+1PO6cDvwfNt8pfN+AIu3de4AfRcQrdd5O0Y+A19P9t7AtufDXPC8D25SlbUMq8JWszYWzkudIBaaSx6vNk/RpSQ9K2iDpKdIdxGI9qCcK758HnoyIlwvT5PW9Oce1WtJTeV3fJl1AAG8EHius6w+VdraKLmCnsrSdSP8Nmg1E6e7focCDwKpe8j9PuvN3TpX5byyd/4XXYBTY/gSI9KjLrBZfAD4h6XWFtJHA6ir5V9P9u38Lkl4DfAi4Nt8AuJ4tH/1OLDv/JxbmfbqQ/mTZcj8uW+6knL5bDzHX6tNl657TS/4/5b+7DnC7Q56742iePwJjST9MJXsAv6uYuw9y/b7PkP4bXBYRr0haT/qR6avHgBeBkWUF0ZLVwO6F6Tf1Yd2/A7aWNC4iHs5p+wLL+hGnWdHVwM9J11Rvj3xLLgf+XVKlOx5/iogxgxVcwWgggKfqsG4bhiLi/lyVZhabfz+eJD2yrWQUqR5sTz5AqtpTavh0DfAzSa+LiD/ntMUR8a4qy/9XRHy+yryjIuJnFdLX9hBzrXrabiWj8991A9zukOc7f81zHfB5SWMkvSo/bnof6T+ugdqRdCH/mVS4+gJb3mGrSUSsBn4KXCBppxzrWyT9Y84yD/hk3o8RpC+kWtf9LOk2/FmSdpD0TmAq6YfbrN8i4g+kH7wjSOdYLcu8RKpDezb9+0epPz4A/CpfC2a1OoNU37pUmLkN2F3SgcVMknYnNS66tZf1TSc96fmjpMdJVYG2AerZ9dbPgA80uP7dB4A1wEMN3GZLcuGvec4C/h9wB7CeVOH8+Ii4fxDWvRC4hXRn7Q/AC3R/NNtX04BtgQdIsV7P5v/YvpO392vgV5T90Er6nKSbe1j3x0mtvdYA3wf+NVdUNhuoGcDBfSxYXQ1sB0ypT0ibuuQYLekMUsOpz9VrWzY8RcRy0g2ET+bp3wHfAq6RNDH3wrA38EPgZ1XuvAEgaTTpKdGRpHrj+5GewJxP5Va/g+WrpJsScyS9uRSLpK9K+pvB3JCkDkmnkgrNpze4nmFrigi//PLLr2HxAlYA76mQvjXp8epYUuv1c3L62Jy+dSHvMTntzDw9CXiFVEe1+PpgL7F0AidWWMezpLpH1wMTm33M/Boar/Jzm1Td5gWgM0+/CvgsqeeH50n/8H8Z2K6X9c4C7qmQ/kZSHfR9gBOAO6osv+l6qhLz82XXzQ1l25hNqr/+DKn7ozOA15StJ4A9K2z3pbJ1P5nnla7r0vW2hvRIe0qzP8dWeSkfKDMzMzNrA37sa2ZmZtZGXPgzM+snbTnk26bhp5odm1mRpGVVztXjmx2bNZ4f+5qZmZm1kWHXz9/IkSNj7NixVec/++yz7LBD63am7/gGppHx3XPPPU9GxOt6z9l6yq+TVv9c+8r70zqG6nXS029JK30erRQLtFY8QyWWplwjzW5xMtivAw44IHpy++239zi/2RzfwDQyPuDuaIFzvj+v8uuk1T/XvvL+tI6hep309FvSSp9HK8US0VrxDJVYmnGNuM6fmZmZWRtx4c/MzMysjbjwZ2ZmZtZGXPgzMzMzayMu/JmZmZm1kWHX1ctwNHbWTZverzjvvU2MxPpD0grSuJUvAxsjYoKkXUkDs48ljX95TESslyTgIuAI4DnghIj4VV7PdODzebXnRMScnH4AaZzL7UnjV56WW5CZ+fujDpau2sAJPq42hPnOn1ljvDsi9ouICXl6FnBrRIwDbs3TAIcD4/LrZOBSgFxYPAM4CDgQOEPSiLzMpcBJheWm1H93zMxsqPKdv2Gk+B8++L/RFjcVmJTfzwE6gc/m9KvynbvFknaRNCrnXRQR6wAkLQKmSOoEdoqIxTn9KuAo4OaG7YmZmQ0pLvwNMS7gDUkB/FRSAN+OiMuAjohYnec/DnTk96OBxwrLrsxpPaWvrJC+BUknk+4m0tHRQWdn56Z5XV1d3aaHOu/PZjPHb9z0fjgdEzPrv7oV/iTtAlwO7EP68ftn4CEGqZ6TJeWFwf4u11Mhspj3yimtMVTOEPOuiFgl6fXAIkm/Lc6MiMgFw7rKhc7LACZMmBCTJk3aNK+zs5Pi9FDn/dmsW9204/u3DjMbXup55+8i4JaIOFrStsBrgM+R6jmdJ2kWqZ7TZ+lez+kgUh2mgwr1nCaQCpD3SJofEevrGHfT9bdAV6/12MBExKr8d42kG0h19p6QNCoiVufHumty9lXA7oXFx+S0VWx+TFxK78zpYyrkNzMzq6guhT9JOwP/AJwAEBEvAS9JGpR6TsD36xG3JS40Dh5JOwCviohn8vvJwFnAfGA6cF7+e2NeZD5wqqS5pH+ENuQC4kLgPwuNPCYDp0fEOklPS5oILAGmAV9v1P6ZmdnQU687f3sAfwa+K2lf4B7gNAavnlM3PdVlKtfqdYG6urqYOf7lhm+3eEyKdYTKDYXj12LxdQA3pJoNbA1cGxG3SLoLmCdpBvAH4JicfwGp+sNyUhWIjwHkQt7ZwF0531mlf4qAj7O5q5ebcWMPMzPrQb0Kf1sD+wOfiIglki5ic1cWwODWc+qpLlO5Vq8L1NnZyQV3PNv4DS8tbrP6aXHllB1a/vi1UnwR8Qiwb4X0tcAhFdIDOKXKumYDsyuk302qW2tmZtarevXztxJYGRFL8vT1pMLgE/lxLn2o51Qp3czMzMz6oS6Fv4h4HHhM0tty0iHAA2yu5wRb1nOapmQiuZ4TsBCYLGlErus0OaeZmZmZWT/Us7XvJ4BrckvfR0h1l17F4NVzMjMzM7M+qlvhLyLuI3XRUm5Q6jmZmZmZWd95bF8zM6s7SbMlrZF0fyHtTEmrJN2XX0cU5p0uabmkhyQdVkifktOW5/5iS+l7SFqS06/LT52Q9Oo8vTzPH9uYPTZrXS78mZlZI1xJ6qe13IURsV9+LQCQtBdwLLB3XuabkraStBVwCWlggL2A43JegPPzuvYE1gMzcvoMYH1OvzDnM2trLvyZmVndRcTPgVrrbE8F5kbEixHxKKk++IH5tTwiHsmDB8wFpuYhQg8m9SwBaRCBowrrKg0Lej1wSM5v1rbq2eDDzMysN6dKmgbcDczMw3eOBhYX8hQ7+C/v+P8gYDfgqYjYWCH/psECImKjpA05/5PlgdQ6YEDH9t07w29mx/Kt1rF9K8XjWKpz4c/MzJrlUuBs0tjtZwMXAP/crGBqHTDg69fcyAVLN/98rji+cr5GaLWO7VspHsdSnR/7mplZU0TEExHxckS8AnyH9FgX+t7x/1pgF0lbl6V3W1eev3POb9a2XPgzM7OmKI34lH0AKLUEng8cm1vq7gGMA+4k9fk6Lrfs3ZbUKGR+7i7sduDovHz5IAKlwQWOBm7L+c3alh/7mplZ3Un6PjAJGClpJXAGMEnSfqTHviuAfwGIiGWS5pFGhtoInBIRL+f1nEoa6WkrYHZELMub+CwwV9I5wL3AFTn9CuBqSctJDU6OrfOumrU8F/6sT5au2sAJs27aNL3ivPc2MRozGyoi4rgKyVdUSCvlPxc4t0L6AtKoUOXpj7D5sXEx/QXgQ30K1myY82NfMzMzszbiwp+ZmZlZG3Hhz8zMzKyNuPBnZmZm1kZc+DMzMzNrI3Ur/ElaIWmppPsk3Z3TdpW0SNLD+e+InC5JF0taLuk3kvYvrGd6zv+wpOnVtmdmZmZmvav3nb93R8R+ETEhT88Cbo2IccCteRrgcFInnuNI4ypeCqmwSOoL6iBSE/4zSgVGMzMzM+u7Rj/2nQrMye/nAEcV0q+KZDFpmJ5RwGHAoohYlwf7XgRMaXDMZgMmaStJ90r6SZ7eQ9KSfLf7ujxaAXlEg+ty+hJJYwvrOD2nPyTpsEL6lJy2XNKs8m2bmZkV1bOT5wB+KimAb+cBszsiYnWe/zjQkd+PBh4rLLsyp1VL70bSyaQ7hnR0dNDZ2Vk1qK6urh7nN1tXVxczx7/c7DCq6tgeZo7fuGm61Y5lC3++pwEPAjvl6fOBCyNirqRvATNId7xnAOsjYk9Jx+Z8H5a0F2lkgr2BNwI/k/TWvK5LgENJ18ddkuZHxAON2jEzMxta6ln4e1dErJL0emCRpN8WZ0ZE5ILhgOWC5WUAEyZMiEmTJlXN29nZSU/zm62zs5ML7ni22WFUNXP8Ri5Yuvm0WXH8pOYFU0Erfr6SxgDvJY1W8ClJAg4GPpKzzAHOJBX+pub3ANcD38j5pwJzI+JF4NE8VFVpNIPleXQDJM3NeV34MzOziupW+IuIVfnvGkk3kH6onpA0KiJW58e6a3L2VcDuhcXH5LRVpLEgi+md9YrZrE6+BnwG2DFP7wY8FRGlW6jFO9qb7nZHxEZJG3L+0cDiwjqLy5TfHT+oUhA93SFv4Tum/eL92ayV79SbWXPUpfAnaQfgVRHxTH4/GTgLmA9MB87Lf2/Mi8wHTs13LQ4CNuQC4kLgPwuNPCYDp9cjZrN6kHQksCYi7pE0qZmx9HSHvBXvmA6E92ezbmNxt9idejNrjnrd+esAbkhPq9gauDYibpF0FzBP0gzgD8AxOf8C4AhgOfAc8DGAiFgn6WzgrpzvrIhYV6eYzerhncD7JR0BbEeq83cRqVHT1vnuX+lON2y+C75S0tbAzsBaqt8dp4d0MzOzLdSl8JfrH+1bIX0tcEiF9ABOqbKu2cDswY7RrBEi4nTy3ep85+/TEXG8pB8ARwNz2fIu+HTgl3n+bbl+7HzgWklfJTX4GAfcCQgYJ2kPUqHvWDbXJTQzM9tCPRt8mFl1nwXmSjoHuBe4IqdfAVydG3SsIxXmiIhlkuaRGnJsBE6JiJcBJJ0KLAS2AmZHxLKG7omZmQ0pLvyZNUhEdJIbLOW74wdWyPMC8KEqy59LajFcnr6AVHXCzMysVx7b18zMzKyNuPBnZmZm1kZc+DMzMzNrIy78mZmZmbURF/7MzMzM2ogLf2ZmZmZtxIU/MzMzszbifv6sbsYWxxQ9771NjMTMzMxKXPizQVMs7A1lLrSamdlw5sKfDchwKfCZmZm1Cxf+rCl6KjQW77b1pXDZ03I93cFzAdbMzNqJC3/WEH0pYLkwZmZmVj91be0raStJ90r6SZ7eQ9ISScslXSdp25z+6jy9PM8fW1jH6Tn9IUmH1TNeG9rGzrqJpas2uPBoZmbWg3p39XIa8GBh+nzgwojYE1gPzMjpM4D1Of3CnA9JewHHAnsDU4BvStqqzjHbMDR21k2bXmZmZu2sbo99JY0B3gucC3xKkoCDgY/kLHOAM4FLgan5PcD1wDdy/qnA3Ih4EXhU0nLgQOCX9Yrbhj8XAM3MrJ3Vs87f14DPADvm6d2ApyJiY55eCYzO70cDjwFExEZJG3L+0cDiwjqLy2wi6WTgZICOjg46OzurBtXV1dXj/Gbr6upi5viXmx1GVR3bw8zxG3vP2CSDHV8rnytmZmb9UZfCn6QjgTURcY+kSfXYRlFEXAZcBjBhwoSYNKn6Jjs7O+lpfrN1dnZywR3PNjuMqmaO38gFS1u3ndBgx7fi+EmDti4zM7NWUK9f8XcC75d0BLAdsBNwEbCLpK3z3b8xwKqcfxWwO7BS0tbAzsDaQnpJcRkzMzMz66O6NPiIiNMjYkxEjCU12LgtIo4HbgeOztmmAzfm9/PzNHn+bREROf3Y3Bp4D2AccGc9YjYzs/qSNFvSGkn3F9J2lbRI0sP574icLkkX594efiNp/8Iy03P+hyVNL6QfIGlpXubiXHe86jbM2lW9W/uW+yyp8cdyUp2+K3L6FcBuOf1TwCyAiFgGzAMeAG4BTomI1q0QZ1ZG0naS7pT0a0nLJH0xpw9at0eSpuS05ZJmNXofzfrgSlLPDUWzgFsjYhxwa54GOJz0D/84Up3uSyEV5IAzgINIDQDPKBTmLgVOKiw3pZdtmLWluhf+IqIzIo7M7x+JiAMjYs+I+FBuxUtEvJCn98zzHyksf25EvCUi3hYRN9c7XrNB9iJwcETsC+wHTJE0kUHq9ih3fXQJ6YdyL+C4nNes5UTEz4F1ZclTSb0/kP8eVUi/KpLFpGpDo4DDgEURsS4i1gOLSNfVKGCniFicnxxdVbauStswa0utW3PfbBjIP0JdeXKb/AoGr9sjgOWlf5gkzc15H6jfXpkNqo6IWJ3fPw505PebeoHISr099JS+skJ6T9voptaeI8p7FWhmrwCt1oNFK8XjWKpz4c+szvLduXuAPUl36X7P4HZ7VP5DeFCVOKr+sLXaF9NAeX82a5VCSi0iIiRFs7ZRa88RX7/mxm69CjSzV4BW68GileJxLNW58GdWZ7me6n6SdgFuAP6qSXFU/WFrtS+mgfL+bHZCoVPzFu266AlJoyJidX50uyanV+vtYRUwqSy9M6ePqZC/p22YtaVGN/gwa1sR8RSpxfs7yN0e5VmVuj2ixm6P3B2SDXXF3h7Ke4GYllv9TgQ25Ee3C4HJkkbkhh6TgYV53tOSJuaqEtOo3KNEcRtmbcmFvxZRGnd26aoNzQ7FBpGk1+U7fkjaHjiUNN71YHV7dBcwLrce3pbUKGR+/ffMrO8kfZ80POfbJK2UNAM4DzhU0sPAe/I0wALgEWA58B3g4wARsQ44m3Tu3wWcldPIeS7Py/weKDUSrLYNs7bkx75m9TUKmJPr/b0KmBcRP5H0ADBX0jnAvXTv9ujq3KBjHakwR0Qsk1Tq9mgjhW6PJJ1KuhuyFTA7d5Fk1nIi4rgqsw6pkDeAU6qsZzYwu0L63cA+FdLXVtqGWbty4c+sjiLiN8DbK6Q/wubWusX0F4APVVnXucC5FdIXkO6SmJmZ9cqPfc3MzMzaiAt/ZmZmZm3EhT8zMzOzNuLCn5mZmVkbceHPzMzMrI248GdmZmbWRlz4MzMzM2sjdSn8SdpO0p2Sfi1pmaQv5vQ9JC2RtFzSdXlEAvKoBdfl9CWSxhbWdXpOf0jSYfWI18zMzKxd1OvO34vAwRGxL7AfMCWPzXg+cGFE7AmsB2bk/DOA9Tn9wpwPSXuRRjjYG5gCfDOPlGBmZmZm/VCXwl8kXXlym/wK4GDg+pw+Bzgqv5+ap8nzD8kDc08F5kbEixHxKGm8xi1GRTAzMzOz2tRteLd8h+4eYE/gEtIg209FxMacZSUwOr8fDTwGEBEbJW0AdsvpiwurLS5T3NbJwMkAHR0ddHZ2Vo2rq6urx/nNMnN8Oiwd229+34raLb5WPFfMzMwGom6Fvzzo/H6SdgFuAP6qjtu6DLgMYMKECTFp0qSqeTs7O+lpfrOcMOsmIBVcLljaukMut1t8K46fNGjrMjMzawV1b+0bEU8BtwPvAHaRVPplHgOsyu9XAbsD5Pk7A2uL6RWWMTMzM7M+qldr39flO35I2h44FHiQVAg8OmebDtyY38/P0+T5t0VE5PRjc2vgPYBxwJ31iNnMzMysHdTr+d0oYE6u9/cqYF5E/ETSA8BcSecA9wJX5PxXAFdLWg6sI7XwJSKWSZoHPABsBE7Jj5PNzMzMrB/qUviLiN8Ab6+Q/ggVWutGxAvAh6qs61zg3MGO0czMzKwdtW7N/SFqbG64AbDivPc2MRIzMzOzLbV14a/WgloxX295+7NtMzMzs0Zpq8Lf2Fk3MXP8xk3dqjRie0W+E2jWN76TbmY2+Ore1YtZO5O0u6TbJT2Qx7k+LafvKmmRpIfz3xE5XZIuzuNZ/0bS/oV1Tc/5H5Y0vZB+gKSleZmL8+g4ZmZmFbnwZ1ZfG4GZEbEXMBE4JY9ZPQu4NSLGAbfmaYDDSV0ajSONWnMppMIicAZwEKnR1BmlAmPOc1JhuSkN2C8zMxuiXPgzq6OIWB0Rv8rvnyH1dzma7uNZl49zfVUeH3sxqWP0UcBhwKKIWBcR64FFwJQ8b6eIWJz7xryqsC4zM7MttFWdP7NmkjSW1AXSEqAjIlbnWY8DHfn9pnGus9J41j2lr6yQXmn7VcfAbvUxr6Fv4yy36v7010D2p7/H0MyGLxf+zBpA0muBHwL/FhFPF6vlRURIinrH0NMY2K0+5jX0bZzlVt2f/hrI/vT3GJrZ8OXHvmZ1JmkbUsHvmoj4UU5+Ij+yJf9dk9OrjWfdU/qYCulmZmYVufBnVke55e0VwIMR8dXCrOJ41uXjXE/LrX4nAhvy4+GFwGRJI3JDj8nAwjzvaUkT87amFdZlZma2BT/2NauvdwIfBZZKui+nfQ44D5gnaQbwB+CYPG8BcASwHHgO+BhARKyTdDZwV853VkSsy+8/DlwJbA/cnF9mZmYVufDXD+541moVEXcA1frdO6RC/gBOqbKu2cDsCul3A/sMIEwzM2sjLvxV4eHXzMzMbDhynT8zMzOzNlKXwl8jhrQyMzMzs76r152/RgxpZWZmZmZ9VJfCX72HtKpHzGZmZmbtoO4NPuo0pFX5NqoOW1U0c/xGOrbvPtxRSfkylfJU0pflinmr5asWX6tot/g8HJZZY0haATwDvAxsjIgJ+enPdcBYYAVwTESsz31aXkTqFuk54ITSDYdcPejzebXnRMScnH4Am7tEWgCcllvXm7Wduhb+GjWkVU/DVhWdMOsmZo7fyAVLK+z20mfLEmo8NH1Yrji00glVWhNXja9FtFt8Hg7LrKHeHRFPFqZLVYXOkzQrT3+W7lWFDiJVFTqoUFVoAhDAPZLm5ydHlwInkW5ELCA9RXKfmNaW6tbat85DWpmZ2fA3KFWF8rydImJxvtt3VWFdZm2nLrdwahjS6jy2HNLqVElzSf/FbYiI1ZIWAv9ZaOQxGTi9HjGbmVlTBfDT/ETo2/mJzmBVFRqd35end1NrFaLy6iXNrB7S1dXVUtVTWikex1JdvZ7fNWJIKzMzGz7eFRGrJL0eWCTpt8WZg1lVqJpaqxB9/Zobu1UvaWb1kM7OTqrF2QytFI9jqa4uhb9GDGllZmbDR0Ssyn/XSLqB1L3XE5JG5SdBtVYVmlSW3pnTx1TIb9aWPMKHmZk1laQdJO1Yek+q4nM/m6sKwZZVhablAQImkqsKAQuByZJG5OpCk4GFed7TkibmaknTCusyazut22zTzMzaRQdwQ+4RYmvg2oi4RdJdDF5VoY+zuauXm3FLX2tjLvyZmVlTRcQjwL4V0tcySFWFIuJuYJ8BB2s2DPixr5mZmVkb8Z2/BhpbpWNnMzMzs0bxnT8zMzOzNuLCn5mZmVkbceHPzMzMrI248GdWR5JmS1oj6f5C2q6SFkl6OP8dkdMl6WJJyyX9RtL+hWWm5/wPS5peSD9A0tK8zMW5DzMzM7OqXPgzq68rgSllabOAWyNiHHBrngY4HBiXXycDl0IqLAJnkMa9PhA4ozDe9aXASYXlyrdlZmbWjQt/ZnUUET8HysejngrMye/nAEcV0q+KZDGwSx7S6jBgUUSsi4j1wCJgSp63U0Qszv2eXVVYl5mZWUXu6sWs8TrycFMAj5NGNwAYDTxWyLcyp/WUvrJCekWSTibdUaSjo4POzs5N87q6urpNt4qZ4zduet+X+Fp1f/prIPvT32NoZsOXC39mTRQRISkatK3LgMsAJgHPoEoAACAASURBVEyYEJMmTdo0r7Ozk9J0sT/KFee9txGhVXVCMZbjJ9W8XHF/hoOB7E9/j6GZDV91eexb70ruZkPcE/mRLfnvmpy+Cti9kG9MTuspfUyFdGszY2fdtOllZtabetX5u5L6VnI3G8rmA6V/ZqYDNxbSp+V/iCYCG/Lj4YXAZEkj8jUwGViY5z0taWJu5TutsK6WVSyouLBiZtZ4dSn81bOSez3iNasXSd8Hfgm8TdJKSTOA84BDJT0MvCdPAywAHgGWA98BPg4QEeuAs4G78uusnEbOc3le5vfAzY3YLzMzG7oaWedvsCq5b6GniuxFM8dvpGP77hWgW43jG5jBjm+gFeQj4rgqsw6pkDeAU6qsZzYwu0L63cA+A4nRzMzaS1MafAx2JfeeKrIXnTDrJmaO38gFS1u3nYvjG5jBjs8V5Iempas2bGro0OxGK2ZmraaRv+JPSBoVEav7UMl9Ull6ZwPiNLMqWqklsJmZ9U8jO3kelEruDYzXzMzMbNipy52/XMl9EjBS0kpSq93zgHm5wvsfgGNy9gXAEaQK688BH4NUyV1SqZI7dK/kbmbDhO8mmpk1Vl0Kf/Wu5G5mZmZm/eOxfc3MzMzaSOs22zQzs038eNzMBosLf2ZmDeRCnJk1mx/7mpmZmbUR3/kzM8vKxxr2nTkzG45858/MzMysjbjwZ2ZmZtZGXPgzMzMzayMu/JmZmZm1ERf+zMzMzNqIW/uamQ2ywejLr7iOmeM3Uvy6Lm+V3J91DiQ2MxvaXPgzs0HXyoWMVo7NzKwRXPgzsyGh2YW2Zm/fzGywuPBnZm1tMB6huiBoZkNJyxf+JE0BLgK2Ai6PiPOaHJJZy/F10lw9FSD7Urjsb0HUeudrxGyzli78SdoKuAQ4FFgJ3CVpfkQ80NzIzFpHPa+TRhdG+ltQ8p23zRrRGGSoPQL3b4lZdy1d+AMOBJZHxCMAkuYCUwFfsGabNeU6GUhhod6xzBzfuG3bkODfErMCRUSzY6hK0tHAlIg4MU9/FDgoIk4ty3cycHKefBvwUA+rHQk8WYdwB4vjG5hGxvfmiHhdg7ZV1SBdJ63+ufaV96d1NP06qcNvSSt9Hq0UC7RWPEMlloZfI61+568mEXEZcFkteSXdHRET6hxSvzm+gWn1+Jqpp+tkuB0374/1R62/Ja30ebRSLNBa8TiW6lp9hI9VwO6F6TE5zcw283Vi1jNfI2YFrV74uwsYJ2kPSdsCxwLzmxyTWavxdWLWM18jZgUt/dg3IjZKOhVYSGqePzsilg1wtTU9Hm4ixzcwrR7foBuk62S4HTfvj21Sh9+SVvo8WikWaK14HEsVLd3gw8zMzMwGV6s/9jUzMzOzQeTCn5mZmVkbGZaFP0mzJa2RdH+V+ZJ0saTlkn4jaf8Wi2+SpA2S7suvLzQ4vt0l3S7pAUnLJJ1WIU9TjmGNsTX1+A01kqZIeih/lrOaHc9A1HJ+DEWStpJ0r6SfNDuWdtfo66XaOS3pTEmrCt9zRxSWOT3H95CkwwY5nhWSluZt3p3TdpW0SNLD+e+InF633wlJbyvs+32Snpb0b408LpV+y/tzLCRNz/kfljR9oHHVJCKG3Qv4B2B/4P4q848AbgYETASWtFh8k4CfNPH4jQL2z+93BH4H7NUKx7DG2Jp6/IbSi1T5/ffA/wdsC/y6/HgOpVct58dQfAGfAq71ed30z6Hh10u1cxo4E/h0hfx75bheDeyR491qEONZAYwsS/syMCu/nwWcn9835Hcify6PA29u5HGp9Fve12MB7Ao8kv+OyO9H1PtcHpZ3/iLi58C6HrJMBa6KZDGwi6RRjYmupviaKiJWR8Sv8vtngAeB0WXZmnIMa4zNardp2KuIeAkoDXs1JA3H80PSGOC9wOXNjsUaf73045yeCsyNiBcj4lFgeY67nqYCc/L7OcBRhfRG/E4cAvw+Iv7QS4yDelyq/Jb39VgcBiyKiHURsR5YBEwZSFy1GJaFvxqMBh4rTK+k9X4g3iHp15JulrR3s4KQNBZ4O7CkbFbTj2EPsUGLHL8hoOmfY730cn4MJV8DPgO80uxArLnXS4Vz+tT8CHF26fFiA2IM4KeS7lEaDg+gIyJW5/ePAx0NiqXkWOD7helmHJeSvh6LppxT7Vr4a3W/Io31ty/wdeDHzQhC0muBHwL/FhFPNyOGanqJrSWOnzVPK5+7fSHpSGBNRNzT7FisuSqc05cCbwH2A1YDFzQolHdFxP7A4cApkv6hODPSs8yG9SGn1Gn3+4Ef5KRmHZctNPpY9EW7Fv5aeqifiHg6Irry+wXANpJGNjIGSduQvmiuiYgfVcjStGPYW2ytcPyGkJa+FvqjhnN3KHkn8H5JK0iPGA+W9L3mhtTWmnK9VDqnI+KJiHg5Il4BvsPmR5h1jTEiVuW/a4Ab8nafKD3OzX/XNCKW7HDgVxHxRI6rKceloK/HoinnVLsW/uYD03Lrm4nAhsJt2qaT9AZJyu8PJH1Oaxu4fQFXAA9GxFerZGvKMawltmYfvyFmWA17VeO5O2RExOkRMSYixpI+m9si4p+aHFY7a/j1Uu2cLqs79wGg1OJ0PnCspFdL2gMYB9w5SLHsIGnH0ntgct7ufKDUSnU6cGMhlnr/ThxH4ZFvM45Lmb4ei4XAZEkj8iPqyTmtrlp6eLf+kvR9UovPkZJWAmcA2wBExLeABaSWN8uB54CPtVh8RwP/Kmkj8DxwbL593CjvBD4KLJV0X077HPCmQozNOoa1xNbs4zdkRH2GUGymiudHvgNsNiBNul6qfecdJ2k/0mPFFcC/5BiXSZoHPABsBE6JiJcHKZYO4Ib8v/XWwLURcYuku4B5kmYAfwCOyfnr+juRC6CHkvc9+3KjjkuV3/Lz6MOxiIh1ks4m/WMBcFZE1L1BqId3MzMzM2sj7frY18zMzKwtufBnZmZm1kZc+DMzMzNrIy78mZmZmbURF/6s7ajCYNy95D9GmwdVv7be8Zm1Al8nZsOXW/ta28k90neRxlncp5e844B5wMERsV7S63PnpmbDmq8Ts+HLd/6s7VQajFvSWyTdkser/IWkv8qzTgIuyQNu4x80axe+TsyGLxf+zJLLgE9ExAHAp4Fv5vS3Am+V9H8lLZY0pWkRmjWfrxOzYWBYjvBh1hd5wPS/A36Qe64HeHX+uzVpGKBJpDEXfy5pfEQ81eg4zZrJ14nZ8OHCn1m6A/5UROxXYd5KYElE/AV4VNLvSD9yd1XIazac+ToxGyb82NfaXkQ8TfrB+hCkgdQl7Ztn/5h0NwNJI0mPtx5pRpxmzeTrxGz4cOHP2k4ejPuXwNskrcwDcB8PzJD0a2AZMDVnXwislfQAcDvw7xGxthlxmzWSrxOz4ctdvZiZmZm1Ed/5MzMzM2sjLvyZmZmZtREX/iqQtELS85K6JD0h6UpJr5XUKemFnP6kpB9JGlW27F6S5kvaIOkZSbdL+rvC/LGSIq+jK29rVmG+JP27pIdzDH+U9CVJry7kGSPphzmGDZLul3RClfWXXh/O86+U9FKO7Zm87Jck7VzDcTlB0st5fU9Luk/SkXneJEmvVNjuO/L80rHbvbC+90haUWE7nZLWF/e5EPs5hembC9v5S96v0vR38/GZVLaO2ZLm9ravw41Sx7xnVUifKulxSd8rO7avzufFH/N5+HA+L1XI0ynpxF622+2zlPStwmf0Uv7cStM393HbPV6LFWI5Nl9vKkvfWmkYsyPzebyybP6Rku6U9KyktZKukTSmMP8ESXf0su0T8nVZug7/vrDfz1a4Zt9Ufnwl7SLp0vx5PSdpqaSPlW1nRd6XHQppJ0rq7Cm+nE+SPpm/E55Vquf3A0nj8/wrc5wHFpbZU1Lk98sK8b9c+Hy6JH2ut+33l8/tTcudKel7henI58irCmnnSLqyMD1D0m+VfguekLRA0o55Xum3okvSOkmLlDv1ztsq7l+XpKcK6616Lqnn7+1vqf/XYEj6TNlyK1X2G1DluP1Fm38TfyfpG5WOuba8jrfPn+H0snxfUOrzsscylqTDJP08b/fPkv5b0vtr2adazrceRYRfZS9gBfCe/H40cD9wHtAJnJjTdwF+ClxTWO4twHrgXGBXYEfgk6Qhkt6R84wFAtg6T78DeA6Ykqe/Djyc07cG9gbuBG4sbOd24GvADjnP24HDK62/wr5dCZyT328H/G1e3/3ADr0clxOAO/L7VwGfyLGPILX0W9nDsp3AWuCyQtp7gBVl+cYCL5NGFvhQtdh72q9C2kfzsdw+Tx8CrAFe1+xzrAnn9HGk1pcqS78euKD8+AHz83m3Tz7HJuZjeXHZZ3piD9us+lnm+WcC36uQ3qdtU+FarBLPdsBTwKSy9COBJ/K2up3HwNHA08BHgO2BNwCzSd8RI8qvix62fXs+/2+qcpy2uGbL9nFb4G5gAbAHsA0wJcf9qcIyK/J2PldIOxHorOEcuRj4PXAwqf++15AaeMzK86/M6/5pYZk9gaiwrh7PDZ/bg3tuV9pmPqfWAh8ppJ0DXJnf/2M+f96ep3cFpgM7Fj7v0m/Fa4BrgMU97V+t51IhX7fPJqdNon/X4FrgyVL8OX0lZdd7T8eNdF3tnc+dPwGjeruOgX8A/gx05Om/BjYAf9XLdkv7dSKwM+k39R+B7/R1n3r7PCq9fOevFxGxCriZdLEW058idW9Q7PPqTOCXEfEfEbEuIp6JiIuBq4Hzq6z/l6RWc/sojY/5ceD4iPhlRGyMiGXAB4Epkg7Oi/0t6QJ+Nue5NyJ6L+lvue0XIuIu4P3AbsDHelmkuOwrpAtwe1KhtxYXA8dJ6in/NGAx6Utheg/5aonxauAh4CxJ2wPfBj4ZEX8eyHqHqB+TPuO/LyVIGkEq+FxVzCjpEGAy8MGIuD+fY4uBfwJOkbRnjdvs82fZn21XuRa3EBEvkMafnVYhzmsjYmNZLCIVHs6JiGsj4vmIeJz0Zd0F/K8a9+nNpC/1k4HDJL2hluXKfBR4E6mg8WhE/CUibiH9c3mWpJ0Keb8CfFrSLrWuPH/3nAIcFxG3RcSLEfFcRFwTEecVss4B/kbSP/ZjH+ql7c/tHnwZ+KKkSn36/i3p9+revK11ETEnIp6pEMdzwLWU/Q5W0odzqVd9vAYfJLVO/1RftlGUr6tlwIdJBbqZhVgqXseRhkGcB3wjx3s58KWI+G0v+/VV4OyIuDwiNkTEKxHx3xFx0mDuUzUu/PVC6THlEcC9Zem7Af8DWF5IPhT4QYXVzAPemQsgxXVI0jtJ/2ncS7oztTIi7izmi4jHSF80h+akxcAlSo+x3tTffSus/xlgEYUvz97kL5PSBfhwjYutAr4DfLGHPNNI/2FeQ7rAOmqNqYr/CfwzMBe4PyLa7pEvQEQ8z5YFn2OA30bEr8uyH0rqsPexsnUsIf3XeUiNm+3PZ9nnbVe5FquZAxxduhaVqju8L6eXexupwNXtms7/+PyQzddjb6YBd0fED0lf5sfXuFzRocDNEfFsWfoPSXc031FIu5t09+jTfVh/xe+eCp4D/pP0dKMl+Nzu0Y9Id5dOqDBvCSn2L0p6p8qq2ZTF8VrSeXtvtTwFtZ5LtejrNfi/gX+TtOtANhoRLwM30v03safr+LOkwvQPSXc6v9LLJt4G7E66w9ibQdmnci78VffjXIfhDuC/SV94ABdL2kC6FTuS9OizZCSwusK6VpOOdfHDe5L0yOBy0q3wW3tYvrSOkfn9h4BfkE6KR5Xq3v1tWf4nJT1VeP11L/v7p7L4qpmYj8vjpMctH4iIDXneG8u2+ZQKdY+yLwHvk7R3+YolvQt4MzAvIu4hPTb4SA0xVRURK4EvkB4x/+tA1jUMlAo+2+XpaVQu9NR6HlY1gM+yL9vu6VqsKCL+L+lR1wdy0jHA7yLiviqxlLbbWyw9mUa6a0L+W37nsRYVj0u+W1na/6IvAJ+Q9Loa179bpfVX8W3gTZIOrzF/I7T9uV1FkH4n/rekbbvNiPgFqWC5P3ATqZ/Gr0raqpDt0/n7fjnwWroXIo8p+66/Paf35VzqTZ+uwXwdLyIVxgaq/Dex6nUcEV2ku50fAGbkwmNPdst/ez1Og7xPm7jwV91REbFLRLw5Ij6e/7uE9NhwZ+BvSHXdxhSWeRKoVDF3FPAKqT5gyciIGBERf50fDfe0fGkdTwJExPqImBURewMdwH2kwmqxIvvIHH/p9WAv+zuaVBjtzeK8vpERMTEiflaY96eybe5SfqciP3L9BrBFBW3S45OfRsSTefpaBvjoN1sGrI+IwfpCGpIi4g7SOXRUfvR+IJu/zIpqOg970d/Psi/b7ula7MlVbP7i/ihljwbLYiltt7dYKsp39vcg3XmGdBzGS+rrY7yKxyXfgR9ZHktE3A/8BJhVvkwVayutv5KIeBE4O79ags/t6iJiAenO4r9UmHdzRLyPVMiZSircFRu6/Ff+Hn9DRLw/In5fmDev7Lv+3Tm95nOpBv25Br8A/OsgPDXa9JtY43W8rOxvT0odoNd6nAZrnzZx4a+fImIpqfLsJYVC189Id+XKHUOqW/FcL6u9DdhdhdZ0sOnR80Tg1gpxPAn8F/BGartzt4V8S/89pLuJjfAV4N3AAYUYticdp39UaqH3OKk+x77aPISUDVyp4PNPwMKIeKJCnp8BB6nQMhtA0kGkRxW39bSBAX6Wfd52lWuxJ1cDhyi1RJ9IenRXyUOkH81u17RSC74PUuF6rGA6IOC+fByWFNL74mfA4RXupH8QeJFUFaTcGcBJpB+x3twKjJE0ocZ4vktqjPA/aszfCD63q/sP4HOkhhdbyPXNbs0x9Fqvrxd9PZd60udrMNe1+xFpn/slr/99bP5NHKzruOQh4DHSPvRqMPapnAt/AzOHdOft/Xn6i8DfSTpX0q6SdpT0CdIXUq+3bCPid8C3gGskTZS0VX48+kPgZ6W7bJLOl7SPUhcVO5IeZy6PPg6npNTtwAGkCsXrSV/odZcrMV8AFJuwH0VqPbcXqXLzfqRWU7+g+2OyrSRtV3h1e5RhvbqKVNA/icqPxcjn2a3ADyXtnc/DicD3gEsjoljHc+uyz2Mbav8sB7rtovJrsadtrCBV5/g+sChXIK+UL0j15j4v6SN5/95AqqqxE3BhIbvKjsN2+RHkMaQK4vsVXp8APqLKlfCruZr0I/gDpe6ctpF0GKkR1ZmFqhfF+JcD15EahfQoH9dvAt9X6kZi27wPx6rQFVUh/0ZS4XJQH0UNUNuf2z1su5PUo8OmwopSVzjHShqh5EBSg4ZK/0j0ZVt9Opd6WVdfrsGiL5IaMNbc6Ak2dfv016TvhjcAXx3k67i4X58iPY7/mKSdJL1K0rskXTaY+1SNC38DEBEvAReR6lSUTvp3AfuSmqGvJpXsD8t1jWpxKunE/h6pMcUtpMrbxf8QXgPcQOq24hFS/ZPyL4an1L0PpmJroc9IeoZ06/kq4B7g7ypUJu+rN5Zts0tStf9sLiJ9iZZMB74bEX+MiMdLL9Ij4uMLF9gs4PnCq8f/1K27XPD5f6Rugub3kPWDpG4NbiGdh98DrmDLukeX0v3z+C61f5YD3XZxv7pdizWYQ7puqj3yLa33OtKj4f9Ful4eILVwf2fZP1t/R/fj8DypK4fngavKjsNsUjcfU2qMtfSo9T2kuwVLSJX4vwr8R0T0VLn8LNJnXYtPkj6jS0jfLb8n1WH6P1Xyf5/Bq9s1YD63e/V5uj8dWk8qKD9MOp++B3wlIqrdCS/34Qrf96/P8/p6LlXVh2uwuMyjpH+Yaj33Pyypi9RFy/y8nQMi4k+kAv+gXMdlMV5PalX8z6T6hU+Q7vLeOEj71COP7WtmZmbWRnznz8zMzKyNuPBn3aj7kDHF17eaHZtZLdR9+Kjiq25DjA0F6j6kXLdXs2Oz2vjc7p9mHbdq15ukmvvUrVtsfuxrZmZm1j5858/MzMysjfS5iXKrGzlyZIwdO7bivGeffZYddhiUhjIty/vYOPfcc8+TEVHrCAotZShfJ45vYBodXy3XidI4xJeT+pgLUgvIh0hd1Ywl9Z5wTESslyRS69cjSMPNnRARv8rrmU5q1QppPNg5Of0A0ji82wMLgNOil8deQ/kaqWaoxg1DN/Za4m7Kb0lEDKvXAQccENXcfvvtVecNF97HxiGN89j0c74/r6F8nTi+gWl0fLVcJ6Sud07M77cl9WX2ZdLQl5C6eDo/vz8CuJnU6e5E0ni5kLoxeST/HZHfj8jz7sx5lZc9vLeYhvI1Us1QjTti6MZeS9zN+C3xY18zM2saSTsD/0Dq746IeClSR/BT2dxZ8xxSf2vk9Kvy7+ZiYBdJo4DDSJ12r4uI9aTxUKfkeTtFxOL8Q3tVYV1mbWnYPfY1M7MhZQ/gz8B3lYZIuwc4DeiIzeNxP04a5QLScHWPFZZfmdN6Sl9ZIX0Lkk4mjeRAR0cHnZ2dFQPu6uqqOq+VDdW4YejG3qpxu/BnZmbNtDWwP/CJiFgi6SLSY95NIiIk1b1rioi4DLgMYMKECTFp0qSK+To7O6k2r5UN1bhh6MbeqnH3+thX0mxJayTdX0g7U9IqSffl1xGFeadLWi7poTz2ZCl9Sk5bXhzfT9Iekpbk9OuUx2pVGnf2upy+RNLYwdppMzNrGSuBlRGxJE9fTyoMPpEf2ZL/rsnzVwG7F5Yfk9N6Sh9TId2sbdVS5+9KKo9dd2FE7JdfCwAk7QUcC+ydl/mm0uDVW5HG+DucNCD2cTkvwPl5XXuSxhqckdNnAOtz+oU5n5mZDSORxkl9TNLbctIhpPFb55PG0iX/LY15Oh+YpmQisCE/Hl4ITJY0QtIIYDKwMM97WtLE3FJ4GlXGTzVrF70+9o2In/fhrttUYG6kQcgflbQcODDPWx4RjwBImgtMlfQgcDDwkZxnDnAmaUDtqfk9pP8EvyFJucJuvyxdtYETZt20aXrFee/t76rMhq3ideJrxBrkE8A1+cnPI8DHSDcn5kmaAfwBOCbnXUBq8buc1NXLxwAiYp2ks4G7cr6zImJdfv9xNnf1cnN+9Zt/S2yoG0idv1MlTQPuBmbm1lWjgcWFPMWKteUVcQ8CdgOeioiNFfJvqrwbERslbcj5nywPpNZKuh3bw8zxGzdNt2IlzIFq1cqlg6kd9tGsnUTEfcCECrMOqZA3gFOqrGc2MLtC+t2kPgTNjP4X/i4FziZ1xnk2cAGpU86mqLWS7tevuZELlm7e5RXHV843lLVq5dLB1A77aGZmVi/96ucvIp6IiJcj4hXgO2x+tNvXirhrSX00bV2W3m1def7OOb+ZmZmZ9VO/Cn+lFljZB4BSS+D5wLG5pe4ewDhSz+p3AeNyy95tSY1C5ufb97cDR+flyyv1lir7Hg3cNpD6fmZmZmZWw2NfSd8HJgEjJa0EzgAmSdqP9Nh3BfAvABGxTNI8UkutjcApEfFyXs+ppNZYWwGzI2JZ3sRngbmSzgHuJffynv9enRuNrCMVGM3MzMxsAGpp7XtcheQrKqSV8p8LnFshfQGplVZ5+iNsfmxcTH8B+FBv8ZmZmZlZ7Ty2r5mZmVkbceHPzMzMrI248GdmZmbWRlz4MzMzM2sjLvyZNUAe4/peST/J03tIWiJpuaTrchdI5G6SrsvpS4pDK0o6Pac/JOmwQvqUnLZc0qxG75uZmQ0tLvyZNcZpwIOF6fOBCyNiT2A9MCOnzwDW5/QLcz4k7UXq7mhvYArwzVyg3Aq4BDgc2As4Luc1MzOryIU/szqTNAZ4L3B5nhZwMHB9zjIHOCq/n5qnyfMPyfmnAnMj4sWIeJQ0qP2B+bU8Ih6JiJeAuTmvmZlZRf0d29fMavc14DPAjnl6N+CpiNiYp1cCo/P70cBjABGxUdKGnH80sLiwzuIyj5WlH1QpCEknAycDdHR00NnZWTHYju1h5vgUWrU8zdTV1dWScZU4PjNrdS78mdWRpCOBNRFxj6RJzYwlIi4DLgOYMGFCTJpUOZyvX3MjFyxNXw0rjq+cp5k6OzupFnsrcHxm1upc+DOrr3cC75d0BLAdsBNwEbCLpK3z3b8xwKqcfxWwO7BS0tbAzsDaQnpJcZlq6WZmZltwnT+zOoqI0yNiTESMJTXYuC0ijgduB47O2aYDN+b38/M0ef5tERE5/djcGngPYBxwJ3AXMC63Ht42b2N+A3bNbFBJWiFpqaT7JN2d03aVtEjSw/nviJwuSRfnFu6/kbR/YT3Tc/6HJU0vpB+Q1788L6vG76VZa3Dhz6w5Pgt8StJyUp2+0njZVwC75fRPAbMAImIZMA94ALgFOCUiXs53Dk8FFpJaE8/Lec2GondHxH4RMSFPzwJujYhxwK15GlLr9nH5dTJwKaTCInAGqd7rgcAZpQJjznNSYbkp9d8ds9bkx75mDRIRnUBnfv8I6cepPM8LwIeqLH8ucG6F9AXAgkEM1axVTAUm5fdzSNfPZ3P6Vfmu+GJJu0galfMuioh1AJIWAVMkdQI7RcTinH4VqYX9zQ3bE7MW4sKfmZm1ggB+KimAb+cGSh0RsTrPfxzoyO83tYrPSq3fe0pfWSG9m/60iIfWbBVfyVBu6T1UY2/VuF34MzOzVvCuiFgl6fXAIkm/Lc6MiMgFw7rpT4t4aM1W8ZUM5ZbeQzX2Vo271zp/kmZLWiPp/kLaVyT9Nle0vUHSLjl9rKTnc4Xd+yR9q7BMxcq2/anQa2Zmw0tErMp/1wA3kKpFPJEf55L/rsnZq7V+7yl9TIV0s7ZUS4OPK9myYuwiYJ+I+Bvgd8DphXm/zxV294uI/1lIr1bZtk8Ves3MbHiRtIOkHUvvgcnA/XRv/V7eKn5avkkwEdiQHw8vBCZLGpFvJEwGFuZ5T0uamG88TCusy6zt9Fr4i4ifA+vK0n5aGJ1gMd3/o9pC/o9tp4hYnCvolirbQvfhrMqHuboqksWkftFGWtdT4AAAIABJREFU1bBPZmY2tHQAd0j6NakLo5si4hbgPOBQSQ8D78nTkBo4PUIa5vA7wMcBckOPs0ldIN0FnFVq/JHzXJ6X+T1u7GFtbDDq/P0zcF1heg9J9wJPA5+PiF/Qc2XbvlboXU2Z4V5Jty9atXLpYGqHfTRrJ7n1+74V0tcCh1RID+CUKuuaDcyukH43sM+AgzUbBgZU+JP0H8BG4JqctBp4U0SslXQA8GNJe9e6vv5W6B3ulXT7olUrlw6mdthHMzOzeul34U/SCcCRwCH5vzAi4kXgxfz+Hkm/B95Kz5Vtn5A0KiJW11ih1/7/9u4/2rKyvvP8+xNAZfwFiLlDgE6RliSLSIJaAyS6MqW2UGC3ZTqGQBwplZZ0Cz06IastTKcholnYK2gHoiRlqLFIE0vij4GlZUiFcNuxOyCghOJHCCWWQ9UgjFQBVkzUMt/5Yz+XOlzPrR/357l3v19rnXX3/u5n7/089559z3c/ez/7SJIkTdO0vuEjyUrgPwBvqKrvDMRfnOSgNv0TdIM1HtrHzbYHekOvJEmSpmmfPX9JPkH31PQjk2yj++qci4Fn0z2LCeDWNrL3F4H3Jfk+8E/Av510s+3HgUPpbrSduNn2cuD6JOcB3wDOavGNwJl0N+d+B3jbTBoqSZKk/Uj+quqcIeFrhsSoqk8Dn55i2dCbbadzQ68kSZKmZ1qXfSVJkrQ4mfxJkiT1iMmfJElSj5j8SZIk9YjJnyRJUo+Y/EmSJPWIyZ8kSVKPmPxJkiT1iMmfJElSj5j8SZIk9YjJnyRJUo+Y/EmSFlySg5J8Ncnn2vxxSW5LsiXJJ5M8q8Wf3ea3tOXLBrZxcYs/kOT0gfjKFtuSZM18t00aNSZ/kqRR8C7g/oH5DwIfrqqXADuB81r8PGBni3+4lSPJCcDZwM8AK4GPtoTyIOAjwBnACcA5razUWyZ/0hxK8pwkX07yN0nuTfI7LW6vhtQkOQZ4PfDHbT7Aa4BPtSLrgTe26VVtnrb8ta38KmBDVX23qr4ObAFObq8tVfVQVX0P2NDKSr118EJXQFrivgu8pqp2JTkE+FKSLwC/QdersSHJH9L1ZlzNQK9GkrPpejV+dVKvxo8Bf5nkJ9s+PgK8DtgG3J7kxqq6bz4bKc3QfwH+A/D8Nv8i4Imq2t3mtwFHt+mjgYcBqmp3kidb+aOBWwe2ObjOw5PipwyrRJLzgfMBxsbGGB8fH1rZsUPhohN3Pz0/VblRs2vXrkVT18kWa91Htd77lfwlWQf8S+Cxqnppix0BfBJYBmwFzqqqne0M7PeBM4HvAG+tqq+0dVYD/7Ft9v1Vtb7FXwF8HDgU2Ai8q6pqqn3MqMXSPKqqAna12UPaq+h6NX6txdcDl9Ilf6vaNHS9Gn8wuVcD+HqSiV4NaL0aAEkmejVM/rQoJJn4bLkzyYqFrEtVrQXWAixfvrxWrBhenauuu4ErNu/5+Nz65uHlRs34+DhTtWnULda6j2q997fn7+PAHwDXDsTWADdX1eXtUtMa4D1091Uc316n0H2gndISuUuA5XQffne2Hoqdrcw7gNvokr+VwBf2sg9p0Wj3HN0JvISul+5rLJJejVE8Yx3VM+kJ1u+AvRJ4Q5IzgecAL6DrQDgsycHtODkG2N7KbweOBbYlORh4IfD4QHzC4DpTxaVe2q/kr6q+OHjvUbMKWNGm1wPjdInZKuDa1uNxa5LDkhzVym6qqh0ASTYBK5OMAy+oqltb/Fq6ezu+sJd9SItGVf0AOCnJYcBngZ9eoHoccK/GKPZojOqZ9ATrd2Cq6mLgYoDW8/ebVfXmJH8GvInuHr3VwA1tlRvb/F+35X/VrhTdCPxpkg/R3RpxPPBlIMDxSY6jS/rOZk+vu9RLM7nnb6yqHmnT3wTG2vTTPRfNRA/F3uLbhsT3to9nWOr3aRyIETyrn3WLtY1V9USSW4Cfx14NaV/eA2xI8n7gq8A1LX4N8Cft1ocddMkcVXVvkuvpbnnYDVzQTrxIciFwE3AQsK6q7p3XlkgjZlYGfLSzrpqNbU1nH0v9Po0DMWpn9XNhMbUxyYuB77fE71C6gRkfBG7BXg3pGapqnO4KD+0+1pOHlPlH4FemWP8DwAeGxDfS3VIkiZklf48mOaqqHmmXdR9r8al6KLaz5xLuRHy8xY8ZUn5v+5AWi6OA9e2+vx8Brq+qzyW5D3s1JEkLYCbJ30QPxeX8cM/FhW3U4SnAky15uwn43SSHt3KnARdX1Y4kTyU5lW7Ax7nAVfvYh7QoVNXdwMuGxO3VkCQtiP191Msn6HrtjkyyjW7U7uXA9UnOA74BnNWKb6R7zMsWuke9vA2gJXmXAbe3cu+bGPwBvJM9j3r5Qnuxl31IkiRpGvZ3tO85Uyx67ZCyBVwwxXbWAeuGxO8AXjok/viwfUiSJGl6/Ho3SZKkHjH5kyRJ6hGTP0mSpB4x+ZMkSeoRkz9JkqQeMfmTJEnqEZM/SZKkHjH5kyRJ6hGTP0mSpB4x+ZMkSeoRkz9JkqQeMfmTJEnqEZM/SdKCSvKcJF9O8jdJ7k3yOy1+XJLbkmxJ8skkz2rxZ7f5LW35soFtXdziDyQ5fSC+ssW2JFkz322URonJnyRpoX0XeE1V/RxwErAyyanAB4EPV9VLgJ3Aea38ecDOFv9wK0eSE4CzgZ8BVgIfTXJQkoOAjwBnACcA57SyUi9NO/lL8lNJ7hp4PZXk3UkuTbJ9IH7mwDoHdEY21VmfJGnpqM6uNntIexXwGuBTLb4eeGObXtXmactfmyQtvqGqvltVXwe2ACe315aqeqiqvgdsaGWlXjp4uitW1QN0Z2i0s6rtwGeBt9Gdqf3eYPlJZ2Q/Bvxlkp9siz8CvA7YBtye5Maquo89Z30bkvwh3dne1dOtsyRpNLXPkTuBl9B9JnwNeKKqdrci24Cj2/TRwMMAVbU7yZPAi1r81oHNDq7z8KT4KUPqcD5wPsDY2Bjj4+ND6zp2KFx04u6n56cqN2p27dq1aOo62WKt+6jWe9rJ3ySvBb5WVd/oTr6GevqMDPh6kokzMmhnZABJNgCrktxPd9b3a63MeuBSTP4kacmpqh8AJyU5jK4j4acXoA5rgbUAy5cvrxUrVgwtd9V1N3DF5j0fn1vfPLzcqBkfH2eqNo26xVr3Ua33bCV/ZwOfGJi/MMm5wB3ARVW1kwM/I3sRU5/1PcNSP1s7EKN6ljGb+tBGqa+q6okktwA/DxyW5OD2OXAM3RUm2s9jgW1JDgZeCDw+EJ8wuM5Ucal3Zpz8tfvw3gBc3EJXA5fR3a9xGXAF8PaZ7mdvlvrZ2oEY1bOM2dSHNkp9kuTFwPdb4nco3W1AHwRuAd5Ed4/eauCGtsqNbf6v2/K/qqpKciPwp0k+RHd70fHAl4EAxyc5ji7pO5s9V5Wk3pmNnr8zgK9U1aMAEz8BknwM+FybPdAzsseZ+qxPkrR0HAWsb/f9/QhwfVV9Lsl9wIYk7we+ClzTyl8D/Em7fWgHXTJHVd2b5HrgPmA3cEG7nEySC4GbgIOAdVV17/w1Txots5H8ncPAJd8kR1XVI232l4B72vQBnZG1s7ipzvokSUtEVd0NvGxI/CH23Bs+GP9H4Fem2NYHgA8MiW8ENs64stISMKPn/CV5Ll33/GcGwv85yeYkdwOvBv4P6M7IgIkzsj+nnZG1Xr2JM7L76c74Js7I3gP8Rju7exF7zvqkRSHJsUluSXJfe3jtu1r8iCSbkjzYfh7e4klyZXu80d1JXj6wrdWt/INJVg/EX9GOuS1t3SlHXUmSNKOev6r6e7qkbDD2lr2UP6AzsqnO+qRFZDfdoKevJHk+cGeSTcBbgZur6vL2bMs1dCc7Z9D1ih9PN/DpauCUJEcAlwDL6e6nvbM9EmlnK/MO4Da642gl8IV5bKMkaRHxGz6kOVRVj1TVV9r0t+l6t4/mmQ+pnfzw2mvbQ29vpbvv9SjgdGBTVe1oCd8mum9BOAp4QVXdWlUFXDuwLUmSfshsPepF0j607x99GV0P3djAvbHfBMba9NMPr20mHnG0t/i2IfFh+z/gRyKN4iN1Rv1RP9ZP0qgz+ZPmQZLnAZ8G3l1VTw3eltcGN9Vc12E6j0QaxcchjfqjfqyfpFHnZV9pjiU5hC7xu66qJgZHPdou2dJ+PtbiUz0SaW/xY4bEJUkayuRPmkNt5O01wP1V9aGBRRMPqYUffnjtuW3U76nAk+3y8E3AaUkObyODTwNuasueSnJq29e5+EgkSdJeeNlXmluvBN4CbE5yV4u9F7gcuD7JecA3gLPaso3AmcAW4DvA2wCqakeSy4DbW7n3VdWONv1O4OPAoXSjfB3pK0maksmfNIeq6kt0DzIf5rVDyhdwwRTbWgesGxK/A3jpDKopSeoRL/tKkiT1iMmfJElSj5j8SZIk9YjJnyRJUo844EPSlJat+fwz5rde/voFqokkabbY8ydJktQjJn+SpAWT5NgktyS5L8m9Sd7V4kck2ZTkwfbz8BZPkiuTbElyd5KXD2xrdSv/YJLVA/FXJNnc1rkyg9+vKPWQyZ8kaSHtBi6qqhOAU4ELkpwArAFurqrjgZvbPMAZwPHtdT5wNXTJInAJcApwMnDJRMLYyrxjYL2V89AuaWTNOPlLsrWdUd2V5I4W84xNkrRPVfVIVX2lTX8buB84GlgFrG/F1gNvbNOrgGurcytwWPt+7NOBTVW1o6p2ApuAlW3ZC6rq1vYQ9WsHtiX10mwN+Hh1VX1rYH7ijO3yJGva/Ht45hnbKXRnY6cMnLEtBwq4M8mN7QCeOGO7je6rr1bi11dJ0pKTZBnwMrr/92Ptu6sBvgmMtemjgYcHVtvWYnuLbxsSH7b/8+l6ExkbG2N8fHxoPccOhYtO3P30/FTlRs2uXbsWTV0nW6x1H9V6z9Vo31XAija9HhinS/6ePmMDbk0ycca2gnbGBpBk4oxtnHbG1uITZ2wmf5K0hCR5HvBp4N1V9dTgRZ6qqiQ113WoqrXAWoDly5fXihUrhpa76robuGLzwMfn5r9/enKUR8SPj48zVZtG3WKt+6jWezaSvwL+oh2Yf9QOnnk9Y1vqZ2sHYlTPMmZTH9oo9UmSQ+gSv+uq6jMt/GiSo6rqkdZJ8FiLbweOHVj9mBbbzp5Oh4n4eIsfM6S81Fuzkfy9qqq2J/lRYFOSvx1cOB9nbNM9W9v65uHlFrNRPcuYTX1oo9QX7T7ua4D7q+pDA4tuBFYDl7efNwzEL0yyge72oSdbgngT8LsDgzxOAy6uqh1JnkpyKt3l5HOBq+a8YdIIm3HyV1Xb28/HknyWbpSVZ2ySpP3xSuAtwOYkd7XYe+mSvuuTnAd8AzirLdsInAlsAb4DvA2gJXmXAbe3cu+buJUIeCfwceBQutuGvHVIvTaj5C/Jc4Efqapvt+nTgPfhGZskaT9U1ZeAqZ7i8Noh5Qu4YIptrQPWDYnfAbx0BtWUlpSZ9vyNAZ9tN+YeDPxpVf15ktvxjE2SJGnkzCj5q6qHgJ8bEn8cz9gkSZJGzlw96kXSErRszeefnh7lR1pIkqbm17tJkiT1iMmfJElSj5j8SZIk9YjJnyRJUo+Y/EmSJPWIyZ80h5KsS/JYknsGYkck2ZTkwfbz8BZPkiuTbElyd5KXD6yzupV/MMnqgfgrkmxu61zZviprXixb8/lnvCRJi4PJnzS3Pg6snBRbA9xcVccDN7d5gDOA49vrfOBq6JJF4BK6b8U5Gbhk4NtwrgbeMbDe5H1JmkeeFGkxMPmT5lBVfRHYMSm8CljfptcDbxyIX1udW4HD2ndjnw5sqqodVbUT2ASsbMteUFW3tgeoXzuwLUmShvIhz9L8G6uqR9r0N+m+JhHgaODhgXLbWmxv8W1D4kMlOZ+uR5GxsTHGx8eHV+5QuOjE3fvZlD2m2t5s27Vr17ztazqsn6RRZ/InLaCqqiQ1T/taC6wFWL58ea1YsWJouauuu4ErNh/4v4atbx6+vdk2Pj7OVHUfBdZP0qgz+ZPm36NJjqqqR9ql28dafDtw7EC5Y1psO7BiUny8xY8ZUn5B7O3+Jr8KTpJGh/f8SfPvRmBixO5q4IaB+Llt1O+pwJPt8vBNwGlJDm8DPU4DbmrLnkpyahvle+7AtiRJGsqeP2kOJfkEXa/dkUm20Y3avRy4Psl5wDeAs1rxjcCZwBbgO8DbAKpqR5LLgNtbufdV1cQgknfSjSg+FPhCe0mSNCWTP2kOVdU5Uyx67ZCyBVwwxXbWAeuGxO8AXjqTOs6HwUvCXgLWZEnWAf8SeKyqXtpiRwCfBJYBW4Gzqmpn6+X+fboTpe8Ab62qr7R1VgP/sW32/VW1vsVfwZ6TpI3Au9rxJvXStC/7Jjk2yS1J7ktyb5J3tfilSbYnuau9zhxY5+L2MNoHkpw+EF/ZYluSrBmIH5fkthb/ZJJnTbe+kqSR9XGW6PMwfeafRtFM7vnbDVxUVScApwIXJDmhLftwVZ3UXhsB2rKzgZ+hO/A+muSgJAcBH6E7oE8AzhnYzgfbtl4C7ATOm0F9JUkjyOdhSvNr2pd9283mj7Tpbye5n708Y4zugN1QVd8Fvp5kC93ZGcCWqnoIIMkGYFXb3muAX2tl1gOX0s7yJElL2rw/D3OpPAtzKov5GY+Lte6jWu9ZuecvyTLgZcBtwCuBC5OcC9xB1zu4k+5gu3VgtcEDcPIBewrwIuCJqto9pPzk/U/rgB3FP8hMjeobbTb1oY2S9piv52EulWdhTmUxP+NxsdZ9VOs94+QvyfOATwPvrqqnklwNXAZU+3kF8PaZ7mdvpnvALvSBOBdG9Y02m/rQxqVs8r1PDgDRFJbk8zClUTCj5C/JIXSJ33VV9RmAqnp0YPnHgM+12akOWKaIP053L8fBrffPA1aS+mPieZiX88PPw7yw3SJ0Cu15mEluAn53YJDHacDF7VFJT7VnZ95G9zzMq+azIRM88dGomHby14bbXwPcX1UfGogfNXCfxi8B97TpG4E/TfIh4MfoRlx9GQhwfJLj6JK7s4Ffa938twBvAjbwzINf0hLkh2M/9fV5mD4CSQtlJj1/rwTeAmxOcleLvZdutO5JdJd9twK/DlBV9ya5HriPbqTwBVX1A4AkF9J9i8FBwLqqurdt7z3AhiTvB75Kl2xKWkJ8BIZ8HqY0v2Yy2vdLdL12k23cyzofAD4wJL5x2HptBPDJk+OSJEmaHr/hQ5KkBeYtD5pPM3nIsyRJkhYZe/4kjSxviJek2WfyJ0nSiNnbiY8nRZopL/tKkiT1iD1/khaFid6Oi07czVu9OV6Sps3kT9Ki50hJLWV7exam731Nh8mfpCXND0f11eT3/sdXPneBaqJRY/InacnxW0PUV773tT96nfw5YkrqH3sC1Vebtz/59P2yvu/7rdfJnyRJfeRJUL+Z/Enqtb1dJvMDUdJSZPInSVPw1hD1he/1fjH5k6T94GUy9cW+Bo343l/8Rj75S7IS+H3gIOCPq+ryudiP/9i1mM3XcaLhBv9/XHTiblYsXFU0BY8RaY+RTv6SHAR8BHgdsA24PcmNVXXfwtZMGh0eJwtjfx+864nkwvMYkZ5ppJM/4GRgS1U9BJBkA7AKmPMD1n/eWkQW7DjRvjmgZCR4jMwiPx8Xv1FP/o4GHh6Y3wacMt+VWEwPzRz2vadLzUQb9/ZPZ2+X8fd1iX8R/mMbieNEB24h/rfMxf+IRXCceIzMkfl6D0/1f3/UP5/35/NqIaSqFroOU0ryJmBlVf2bNv8W4JSqunBSufOB89vsTwEPTLHJI4FvzVF1R4VtnD8/XlUvXuhK9PA4sX4zM9/1W/DjpIfHyFQWa71h8dZ9f+o978fIqPf8bQeOHZg/psWeoarWAmv3tbEkd1TV8tmr3uixjb3Uq+PE+s3MqNdvjvTqGJnKYq03LN66j2q9f2ShK7APtwPHJzkuybOAs4EbF7hO0qjxOJH2zmNEGjDSPX9VtTvJhcBNdMPz11XVvQtcLWmkeJxIe+cxIj3TSCd/AFW1Edg4S5vbZ3f+EmAbe6hnx4n1m5lRr9+c6NkxMpXFWm9YvHUfyXqP9IAPSZIkza5Rv+dPkiRJs6gXyV+SlUkeSLIlyZqFrs8wSdYleSzJPQOxI5JsSvJg+3l4iyfJla09dyd5+cA6q1v5B5OsHoi/Isnmts6VSbK3fcxRG49NckuS+5Lcm+RdS7Gdi9WoHCdJtra/4V1J7mixA36PzGJ95vTYnKP6XZpke/sd3pXkzIFlF7f6PZDk9IH4SPz9R92o/Z5G/f25l3rP+efBHNX7OUm+nORvWr1/p8WPS3Jbq98n0w0sIsmz2/yWtnzZwLaGHovzoqqW9Ivu5t6vAT8BPAv4G+CEha7XkHr+IvBy4J6B2H8G1rTpNcAH2/SZwBeAAKcCt7X4EcBD7efhbfrwtuzLrWzaumfsbR9z1MajgJe36ecDfwecsNTauRhfo3ScAFuBIyfFDug9Msv1mdNjc47qdynwm0PKntD+ts8Gjmt/84NG6e8/yq9R/D2N+vtzL/We88+DOap3gOe16UOA21p9rgfObvE/BP5dm34n8Idt+mzgk2166LE4X++bPvT8Pf21PlX1PWDia31GSlV9EdgxKbwKWN+m1wNvHIhfW51bgcOSHAWcDmyqqh1VtRPYBKxsy15QVbdW9667dtK2hu1j1lXVI1X1lTb9beB+uifvL6l2LlKjfpwc6Htk1szlsTmH9ZvKKmBDVX23qr4ObKH724/6339UjNzvadTfn3up95x+HsxhvauqdrXZQ9qrgNcAn5qi3hPt+RTw2iRh6mNxXvQh+Rv2tT5HL1BdDtRYVT3Spr8JjLXpqdq0t/i2IfG97WNOta7vl9GdNS3Zdi4io3ScFPAXSe5M940LcODvkbk2W+/ZuXRhuzy2Lntucxil+i1Gi+X3tBjen0+bo8+DOZPkoCR3AY/RJZtfA56oqt1D6vB0/dryJ4EXLUS9B/Uh+VsSWk/WnA7Nno99ACR5HvBp4N1V9dR812G+2qlpe1VVvRw4A7ggyS8OLhy1v9+o1ae5GvjnwEnAI8AVC1sdLZQRfX8+baE/D6ajqn5QVSfRfVPMycBPL3CVDlgfkr/9+lqfEfXoxCWs9vOxFp+qTXuLHzMkvrd9zIkkh9Ad6NdV1Wf2UYdF285FaGSOk6ra3n4+BnyW7p/rgb5H5tpsvWfnRFU92j6g/gn4GHsuJ41E/RaxxfJ7Gun354Q5/jyYc1X1BHAL8PN0l6Ennp08WIen69eWvxB4nAV+L/Uh+VvMX+tzIzAxcmk1cMNA/Nw2+ulU4MnWTX4TcFqSw9tlntOAm9qyp5Kc2u41OHfStobtY9a1fV8D3F9VHxpYtKTauUiNxHGS5LlJnj8xTfe3vYcDf4/MtVl5z85V5Sbd9/hLdL/Difqd3UYgHgccTzdIaiT+/ovAYvk9jfT7E+b+82AO6/3iJIe16UOB19Hdr3gL8KYp6j3RnjcBf9V6NKc6FufH5BEgS/FFN0ro7+iuy//WQtdnijp+gu7yzPfprv2fR3dfwM3Ag8BfAke0sgE+0tqzGVg+sJ230904ugV420B8Od0HwNeAP2DPA76H7mOO2vgqui78u4G72uvMpdbOxfoaheOEbhTl37TXvRP1mM57ZBbrNKfH5hzV70/a/u+m+5A5aqD8b7X6PUAbDT8qf//F8Bq139Oovz/3Uu85/zyYo3r/LPDVVu97gP/U4j9Bl7xtAf4MeHaLP6fNb2nLf2JgW0OPxfl4+Q0fkiRJPdKHy76SJElqTP4kSZJ6xORPkiSpR0z+JEmSesTkT5IkqUdM/iRJknrE5E+SJKlHTP4kSZJ6xORPkiSpR0z+JEmSesTkT5IkqUdM/iRJknrE5E+SJKlHTP4kSZJ6xORPkiSpR0z+JEmSesTkbx4k+a9J/s9Jsf81yeNJ/ijJ95PsGng9MalskjyU5L4h2x5P8o9tvW8l+UySo/azXickuTHJk0m+neSWJL8wqcyzklya5MEkf59ka5J1SZYNlDk9yRfbNv6/JP8tyRvaskuT/Nch+64kL5lpGyRJ0oEx+Zsf7wLOSPI6gCTPAT4GXAQ8Anyyqp438Dps0vq/CPwo8BNJ/pch27+wqp4HvAR4HvB7+6pQkn8O/HdgM3Ac8GPAZ4G/SPLzA0U/BbwB+DXghcDPAXcCr23beRPwZ8C1wDHAGPCfgH+1rzrMtA2SJOnAmfzNg6p6HPj3wNokzwUuAb5WVR/fz02sBm4ANrbpqfbzBPB/ASftxzYvBf66qn6rqnZU1ber6krgT4APAiT5F8DrgFVVdXtV7a6qJ6vqI1V1TZIAHwIuq6o/bsv+qar+W1W9Yz/bNpM2SJKkA2TyN0+q6s+ArwCfAM5vr31K8j8BbwKua6+zkzxrirIvAv41sGU/Nv06uh67ya4HXpnkUOBfAF+uqoen2MZPAcfS9Q7OigNsgyRJOkAmf/PrncBrgPdNSqjOSvLEwOuWgWX/Gvgu8BfA54FDgNdP2u6VSZ4EvgUcSdfLuC9H0l1ynuwRuvfFEcCLpigz4UUD68zUdNogSZIOkMnfPKqqR+mSm3snLbq+qg4beL16YNnqtnx3Vf0j8Gl++NLv/15VLwR+Fjic7t67ffkWMGxQxVHAPwE7gcenKDPh8YF1prKbLmF9WpKJ+e8PhKfTBkmSdIBM/kZYkmPoegr/tyTfTPJNukvAZyY5cnL5qtoMvB/4SLsfb2/+EviVIfGz6O4F/E4rc3KrxzAPAA8Dv7yX/fw/wLJJsePoksLtkwsfYBskSdIBMvkbbW8B/o7u3rqT2usngW3AOVOss55uxO0b9rHt3wF+IckHkhyR5PlJ/j1wLvAegKr6S2AT8Nkkr0hycCv3b5O8vaoK+A3gt5O8LckLkvxIklclWdv28+fATyd5S5JDkhwB/C7w6araPcM2SJIpCMHZAAASS0lEQVSkA2TyNxp+ddJz/nYl+VG6y7sfrapvDr6AP2SKUb9V9T3g94Hf3tsOq+pB4FV0j27ZSnff3i8Dp1fVfx8o+ia6UcafBJ4E7gGW0/UKUlWfAn4VeDvw/wKP0vXc3dCWPwacAfw68Fhb/wng3+2lbvvVBkmSdODSdd5IkiSpD+z5kyRJ6hGTvyUsyReGXE7eleS9C103SZK0MLzsK0mS1CMHL3QFZtuRRx5Zy5YtG7rs7//+73nuc587vxVaILZ17t15553fqqoXz/uOJUmagSWX/C1btow77rhj6LLx8XFWrFgxvxVaILZ17iX5xrzvVJKkGfKeP0mSpB4x+ZMkSeoRkz9JkqQeMfmTJEnqEZM/SZKkHllyo333ZvP2J3nrms8/Pb/18tcvYG0kSZLmnz1/kiRJPWLyJ0mS1CMmf5IkST0y7eQvybFJbklyX5J7k7yrxS9Nsj3JXe115sA6FyfZkuSBJKcPxFe22JYkawbixyW5rcU/meRZ062vJEmSZtbztxu4qKpOAE4FLkhyQlv24ao6qb02ArRlZwM/A6wEPprkoCQHAR8BzgBOAM4Z2M4H27ZeAuwEzptBfSVJknpv2slfVT1SVV9p098G7geO3ssqq4ANVfXdqvo6sAU4ub22VNVDVfU9YAOwKkmA1wCfauuvB9443fpKkiRplh71kmQZ8DLgNuCVwIVJzgXuoOsd3EmXGN46sNo29iSLD0+KnwK8CHiiqnYPKT95/+cD5wOMjY0xPj4+tJ5jh8JFJ+5+en6qckvBrl27lnT7BvWprZIkzdSMk78kzwM+Dby7qp5KcjVwGVDt5xXA22e6n72pqrXAWoDly5fXihUrhpa76robuGLzniZvffPwckvB+Pg4U/0elpo+tVWSpJmaUfKX5BC6xO+6qvoMQFU9OrD8Y8Dn2ux24NiB1Y9pMaaIPw4cluTg1vs3WF6SJEnTMJPRvgGuAe6vqg8NxI8aKPZLwD1t+kbg7CTPTnIccDzwZeB24Pg2svdZdINCbqyqAm4B3tTWXw3cMN36SpIkaWY9f68E3gJsTnJXi72XbrTuSXSXfbcCvw5QVfcmuR64j26k8AVV9QOAJBcCNwEHAeuq6t62vfcAG5K8H/gqXbIpSZKkaZp28ldVXwIyZNHGvazzAeADQ+Ibh61XVQ/RjQaWJEnSLPAbPiRJknrE5E+SJKlHTP4kSZJ6ZFYe8rwULVvz+aent17++gWsiSRJ0uyx50+SJKlHTP4kSZJ6xORPkiSpR0z+JEmSesTkT5IkqUdM/iRJknrE5E+SJKlHTP4kSZJ6xORPkiSpR0z+JEmSesTkT5IkqUdM/iRJknrE5E+SJKlHTP4kSZJ6ZNrJX5Jjk9yS5L4k9yZ5V4sfkWRTkgfbz8NbPEmuTLIlyd1JXj6wrdWt/INJVg/EX5Fkc1vnyiSZSWMlSZL6biY9f7uBi6rqBOBU4IIkJwBrgJur6njg5jYPcAZwfHudD1wNXbIIXAKcApwMXDKRMLYy7xhYb+UM6itJktR7007+quqRqvpKm/42cD9wNLAKWN+KrQfe2KZXAddW51bgsCRHAacDm6pqR1XtBDYBK9uyF1TVrVVVwLUD25IkSdI0HDwbG0myDHgZcBswVlWPtEXfBMba9NHAwwOrbWuxvcW3DYkP2//5dL2JjI2NMT4+PrSeY4fCRSfufnp+qnKw/+VG1a5duxZlvaejT22VJGmmZpz8JXke8Gng3VX11OBteVVVSWqm+9iXqloLrAVYvnx5rVixYmi5q667gSs272ny1jcPLwfw1jWf369yywbKAWy9/PX7rvA8GB8fZ6rfw1LTp7ZKkjRTMxrtm+QQusTvuqr6TAs/2i7Z0n4+1uLbgWMHVj+mxfYWP2ZIXJIkSdM0k9G+Aa4B7q+qDw0suhGYGLG7GrhhIH5uG/V7KvBkuzx8E3BaksPbQI/TgJvasqeSnNr2de7AtiRJkjQNM7ns+0rgLcDmJHe12HuBy4Hrk5wHfAM4qy3bCJwJbAG+A7wNoKp2JLkMuL2Ve19V7WjT7wQ+DhwKfKG9JEmSNE3TTv6q6kvAVM/de+2Q8gVcMMW21gHrhsTvAF463TpKkiTpmfyGD0mSpB4x+ZMkSeoRkz9JkqQeMfmTJEnqEZM/SZKkHjH5kyRJ6hGTP0mSpB4x+ZMkSeoRkz9JkqQeMfmTJEnqEZM/SZKkHjH5kyRJ6hGTP0mSpB4x+ZMkSeoRkz9JkqQeMfmTJEnqEZM/SZKkHjH5kyRJ6pGDF7oCfbVszeefMb/18tcvUE0kSVKfzKjnL8m6JI8luWcgdmmS7Unuaq8zB5ZdnGRLkgeSnD4QX9liW5KsGYgfl+S2Fv9kkmfNpL6SJEl9N9PLvh8HVg6Jf7iqTmqvjQBJTgDOBn6mrfPRJAclOQj4CHAGcAJwTisL8MG2rZcAO4HzZlhfSZKkXpvRZd+q+mKSZftZfBWwoaq+C3w9yRbg5LZsS1U9BJBkA7Aqyf3Aa4Bfa2XWA5cCV8+kzovR4CViLw9LkqSZmKt7/i5Mci5wB3BRVe0EjgZuHSizrcUAHp4UPwV4EfBEVe0eUv4ZkpwPnA8wNjbG+Pj40EqNHQoXnbj76fmpysH0yu2r7HTX29+6DNq1a9d+l13s+tRWSZJmai6Sv6uBy4BqP68A3j4H+3laVa0F1gIsX768VqxYMbTcVdfdwBWb9zR565uHlwN462Bv236W21fZ6a63v3UZND4+zlS/h6WmT22VJGmmZj35q6pHJ6aTfAz4XJvdDhw7UPSYFmOK+OPAYUkObr1/g+UlSZI0DbP+nL8kRw3M/hIwMRL4RuDsJM9OchxwPPBl4Hbg+Day91l0g0JurKoCbgHe1NZfDdww2/WVJEnqkxn1/CX5BLACODLJNuASYEWSk+gu+24Ffh2gqu5Ncj1wH7AbuKCqftC2cyFwE3AQsK6q7m27eA+wIcn7ga8C18ykvpIkSX0309G+5wwJT5mgVdUHgA8MiW8ENg6JP8SeEcGSJEmaIb/eTZIkqUdM/iRJknrE5E+SJKlHTP4kSZJ6xORPkiSpR0z+JEmSesTkT5IkqUdM/iRJknrE5E+SJKlHTP4kSZJ6xORPkiSpR0z+JEmSesTkT5IkqUdM/iRJknrE5E+SJKlHTP4kSZJ65OCFroBmz7I1n396+qITd7Ni4aoiSZJGlD1/kiRJPTKj5C/JuiSPJblnIHZEkk1JHmw/D2/xJLkyyZYkdyd5+cA6q1v5B5OsHoi/Isnmts6VSTKT+kqSJPXdTHv+Pg6snBRbA9xcVccDN7d5gDOA49vrfOBq6JJF4BLgFOBk4JKJhLGVecfAepP3JUmSpAMwo+Svqr4I7JgUXgWsb9PrgTcOxK+tzq3AYUmOAk4HNlXVjqraCWwCVrZlL6iqW6uqgGsHtiVJkqRpmIsBH2NV9Uib/iYw1qaPBh4eKLetxfYW3zYk/kOSnE/Xm8jY2Bjj4+PDK3ZoNxBiwlTlYHrl9lV2uutNpy5jh+5/XRa7Xbt29aatkiTN1JyO9q2qSlJzuY+2n7XAWoDly5fXihUrhpa76robuGLzniZvffPwcgBvHRg5u7/l9lV2uutNpy4Xnbibs6b4PcyXwdHHWy9//ZytNz4+zlR/c0mS9ExzMdr30XbJlvbzsRbfDhw7UO6YFttb/JghcUmSJE3TXPT83QisBi5vP28YiF+YZAPd4I4nq+qRJDcBvzswyOM04OKq2pHkqSSnArcB5wJXzUF9dQCWTe6xPIAePUmStPBmlPwl+QSwAjgyyTa6UbuXA9cnOQ/4BnBWK74ROBPYAnwHeBtAS/IuA25v5d5XVRODSN5JN6L4UOAL7SVJkqRpmlHyV1XnTLHotUPKFnDBFNtZB6wbEr8DeOlM6ihJkqQ9/IYPSZKkHjH5kyRJ6hGTP0mSpB4x+ZMkSeqROX3IsxaP6T6QWZIkLS72/EmSJPWIyZ8kSVKPeNlXveBlbUmSOvb8SZIk9YjJnyRJUo+Y/GnRWrbm8yxb83k2b3/yGZd1JUnS1Ez+JEmSesQBH9IkDg6RJC1l9vxJkiT1iMmfJElSj5j8SZIk9YjJnyRJUo+Y/EmSJPXInCV/SbYm2ZzkriR3tNgRSTYlebD9PLzFk+TKJFuS3J3k5QPbWd3KP5hk9VzVV5IkqQ/muufv1VV1UlUtb/NrgJur6njg5jYPcAZwfHudD1wNXbIIXAKcApwMXDKRMEqSJOnAzfdl31XA+ja9HnjjQPza6twKHJbkKOB0YFNV7aiqncAmYOU811mSJGnJSFXNzYaTrwM7gQL+qKrWJnmiqg5rywPsrKrDknwOuLyqvtSW3Qy8B1gBPKeq3t/ivw38Q1X93qR9nU/XY8jY2NgrNmzYMLROj+14kkf/Yc/8iUe/cMr6b97+5AGX21fZ6a43nbqMHQo/esT+1WW6+ziQsvv7e5lOXcYOhUf/Yfbqsr9lX/3qV9850KstSdKiMJff8PGqqtqe5EeBTUn+dnBhVVWSWck8q2otsBZg+fLltWLFiqHlrrruBq7YvKfJW988vBzAWwe/5WE/y+2r7HTXm05dLjpxN2dN8XuYrX0cSNn9/b1Mpy4XnbibKzYfPGt1mW69JUlaDObssm9VbW8/HwM+S3fP3qPtci7t52Ot+Hbg2IHVj2mxqeKSJEmahjlJ/pI8N8nzJ6aB04B7gBuBiRG7q4Eb2vSNwLlt1O+pwJNV9QhwE3BaksPbQI/TWkySJEnTMFeXfceAz3a39XEw8KdV9edJbgeuT3Ie8A3grFZ+I3AmsAX4DvA2gKrakeQy4PZW7n1VtWOO6ixJkrTkzUnyV1UPAT83JP448Noh8QIumGJb64B1s11HSZKkPvIbPiRJknrE5E+SJKlHTP4kSZJ6xORPkiSpR0z+JEmSesTkT5IkqUdM/iRJknrE5E+SJKlHTP4kSZJ6xORPkiSpR0z+JEmSesTkT5IkqUdM/iRJknrE5E+SJKlHTP4kSZJ6xORPkiSpR0z+JEmSesTkT5IkqUdM/iRJknpk5JO/JCuTPJBkS5I1C10fSZKkxWykk78kBwEfAc4ATgDOSXLCwtZKkiRp8Rrp5A84GdhSVQ9V1feADcCqBa6TJEnSopWqWug6TCnJm4CVVfVv2vxbgFOq6sJJ5c4Hzm+zPwU8MMUmjwS+NUfVHTW2de79eFW9eAH2K0nStB280BWYDVW1Fli7r3JJ7qiq5fNQpQVnWyVJ0jCjftl3O3DswPwxLSZJkqRpGPXk73bg+CTHJXkWcDZw4wLXSZIkadEa6cu+VbU7yYXATcBBwLqquncGm9znpeElxLZKkqQfMtIDPiRJkjS7Rv2yryRJkmaRyZ8kSVKPLOrkL8mxSW5Jcl+Se5O8q8WPSLIpyYPt5+Et/tNJ/jrJd5P85qRtjfTXyM1yW7cm2ZzkriR3LER79mYabX1zkrtbm/5Hkp8b2NZI/10lSZpvi/qevyRHAUdV1VeSPB+4E3gj8FZgR1Vd3j7wD6+q9yT5UeDHW5mdVfV7bTsHAX8HvA7YRjfK+Jyqum/eGzWF2Wpr29ZWYHlVjeRDoKfR1l8A7q+qnUnOAC6tqlMWw99VkqT5tqh7/qrqkar6Spv+NnA/cDTdV8Ctb8XW0yUOVNVjVXU78P1Jmxr5r5GbxbaOvGm09X9U1c4Wv5XueZCwCP6ukiTNt0Wd/A1Ksgx4GXAbMFZVj7RF3wTG9rH60cDDA/PbWmwkzbCtAAX8RZI721fjjaxptPU84AttelH9XSVJmg8j/Zy//ZXkecCngXdX1VNJnl5WVZVk8V7bnmSW2vqqqtreLg1vSvK3VfXFOarytB1oW5O8mi75e9W8VlSSpEVk0ff8JTmELkG4rqo+08KPtvvGJu4fe2wfm1kUXyM3S22lqra3n48Bn6W7PDpSDrStSX4W+GNgVVU93sKL4u8qSdJ8WtTJX7quoGvobvb/0MCiG4HVbXo1cMM+NjXyXyM3W21N8tw2iIIkzwVOA+6Z/RpP34G2Nck/Az4DvKWq/m6g/Mj/XSVJmm+LfbTvq4D/G9gM/FMLv5fu/rDrgX8GfAM4q6p2JPmfgTuAF7Tyu4AT2iXFM4H/wp6vkfvAvDZmH2arrcCRdL190F32/9Ml0NY/Bn65xQB2V9Xytq2R/rtKkjTfFnXyJ0mSpAOzqC/7SpIk6cCY/EmSJPWIyZ8kSVKPmPxJkiT1iMmfJElSj5j8SZIk9YjJnyRJUo/8/87G51stdNT0AAAAAElFTkSuQmCC\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "## Process outlier data"
      ],
      "metadata": {
        "id": "y_vImwZqrmOd"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "X_outliers = violations_outliers[FEATURES]\n",
        "y_outliers = violations_outliers[TARGETS]\n",
        "\n",
        "# Instantiate encoder/scaler\n",
        "scaler = StandardScaler()\n",
        "ohe = OneHotEncoder(sparse=False)\n",
        "\n",
        "# Scale and Encode Separate Columns\n",
        "scaled_columns_outliers  = scaler.fit_transform(X_outliers[columns_to_scale])\n",
        "encoded_columns_outliers = ohe.fit_transform(X_outliers[columns_to_encode])\n",
        "\n",
        "# Concatenate (Column-Bind) Processed Columns Back Together\n",
        "X_pre_outliers = np.concatenate([scaled_columns_outliers, encoded_columns_outliers], axis=1)\n",
        "np.nan_to_num(X_pre_outliers, copy=False)\n",
        "\n",
        "X_train_outliers, X_test_outliers, y_train_outliers, y_test_outliers = train_test_split(X_pre_outliers, y_outliers, test_size = 0.25, random_state = 0)"
      ],
      "metadata": {
        "id": "e1yYuaEDroxS"
      },
      "execution_count": 354,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "## Outlier Linear Regression"
      ],
      "metadata": {
        "id": "AvZEqFvWrgKh"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "lin_reg_outliers = LinearRegression().fit(X_train_outliers, y_train_outliers)"
      ],
      "metadata": {
        "id": "aTSmMLB_r-cO"
      },
      "execution_count": 355,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "lin_reg_outliers_y_pred_train = lin_reg_outliers.predict(X_train_outliers)\n",
        "lin_reg_outliers_y_pred_test = lin_reg_outliers.predict(X_test_outliers)\n",
        "\n",
        "lin_reg_outliers_r_squared_train = r2_score(y_train_outliers, lin_reg_outliers_y_pred_train)\n",
        "lin_reg_outliers_r_squared_test = r2_score(y_test_outliers, lin_reg_outliers_y_pred_test)\n",
        "\n",
        "lin_reg_outliers_mae_train = mean_absolute_error(y_train_outliers, lin_reg_outliers_y_pred_train)\n",
        "lin_reg_outliers_mae_test = mean_absolute_error(y_test_outliers, lin_reg_outliers_y_pred_test)\n",
        "\n",
        "print('Outlier Simple Linear Regression Scores:')\n",
        "print('Training R^2 = %.3f' % lin_reg_outliers_r_squared_train)\n",
        "print('Training MAE = %.3f' % lin_reg_outliers_mae_train)\n",
        "\n",
        "print('\\nTesting R^2 = %.3f' % lin_reg_outliers_r_squared_test)\n",
        "print('Testing MAE = %.3f' % lin_reg_outliers_mae_test)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 0
        },
        "id": "usGZKXHBsASw",
        "outputId": "0dc68e6a-f7e5-4e5c-e0da-ca5817ea8cd8"
      },
      "execution_count": 356,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Outlier Simple Linear Regression Scores:\n",
            "Training R^2 = 0.275\n",
            "Training MAE = 3416.393\n",
            "\n",
            "Testing R^2 = 0.285\n",
            "Testing MAE = 3416.473\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "## Outlier RANSAC Regression"
      ],
      "metadata": {
        "id": "3M3z-DJLsd_L"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "ransac_reg_outliers = RANSACRegressor(random_state=2).fit(X_train_outliers, y_train_outliers)"
      ],
      "metadata": {
        "id": "Fi_BMX9qsiNT"
      },
      "execution_count": 357,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "ransac_reg_outliers_y_pred_train = ransac_reg_outliers.predict(X_train_outliers)\n",
        "ransac_reg_outliers_y_pred_test = ransac_reg_outliers.predict(X_test_outliers)\n",
        "\n",
        "ransac_reg_outliers_r_squared_train = r2_score(y_train_outliers, ransac_reg_outliers_y_pred_train)\n",
        "ransac_reg_outliers_r_squared_test = r2_score(y_test_outliers, ransac_reg_outliers_y_pred_test)\n",
        "\n",
        "ransac_reg_outliers_mae_train = mean_absolute_error(y_train_outliers, ransac_reg_outliers_y_pred_train)\n",
        "ransac_reg_outliers_mae_test = mean_absolute_error(y_test_outliers, ransac_reg_outliers_y_pred_test)\n",
        "\n",
        "print('Outlier RANSAC Regression Scores:')\n",
        "print('Training R^2 = %.3f' % ransac_reg_outliers_r_squared_train)\n",
        "print('Training MAE = %.3f' % ransac_reg_outliers_mae_train)\n",
        "\n",
        "print('\\nTesting R^2 = %.3f' % ransac_reg_outliers_r_squared_test)\n",
        "print('Testing MAE = %.3f' % ransac_reg_outliers_mae_test)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 0
        },
        "id": "UXhq1BHAsk4h",
        "outputId": "623284f3-9a13-441f-d074-8c7da8002f07"
      },
      "execution_count": 358,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Outlier RANSAC Regression Scores:\n",
            "Training R^2 = 0.035\n",
            "Training MAE = 3127.516\n",
            "\n",
            "Testing R^2 = 0.025\n",
            "Testing MAE = 3135.479\n"
          ]
        }
      ]
    }
  ]
}