NaiveBayes_Classifier.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "naive_bayes_ml.ipynb",
      "provenance": [],
      "collapsed_sections": []
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "accelerator": "GPU"
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "bVfzVfv3K2KD"
      },
      "source": [
        "### Naive Bayes Classification for Face-Mask Detection\n",
        "\n"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Z30YX67YLI08"
      },
      "source": [
        "#### Import Libraries"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "njsvHYwSTj5-"
      },
      "source": [
        "# Import Libraries\n",
        "import numpy as np \n",
        "\n",
        "import sklearn\n",
        "import sklearn.naive_bayes\n",
        "from sklearn.model_selection import train_test_split\n",
        "\n",
        "import pickle\n",
        "import cv2\n",
        "from cv2 import cv2\n",
        "\n",
        "import matplotlib.pyplot as plt\n",
        "import seaborn as sns\n",
        "sns.set()"
      ],
      "execution_count": 1,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "24V4sfG7LARf"
      },
      "source": [
        "#### EDA\n"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "rZES7sgcWBCm"
      },
      "source": [
        "x_path = '/content/drive/MyDrive/Face Detection Dataset/data_pickle/X.pickle'\n",
        "y_path = '/content/drive/MyDrive/Face Detection Dataset/data_pickle/y.pickle'\n",
        "\n",
        "X = pickle.load(open(x_path, 'rb'))\n",
        "y = pickle.load(open(y_path, 'rb'))\n",
        "\n",
        "CATEGORIES = ['Without Mask', 'Incorrectly Wearing Mask', 'Wearing Mask']"
      ],
      "execution_count": 2,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "5SN-iQc-WAFO",
        "outputId": "bc71cdf5-adc9-4a69-9aa9-63ed2737e83c"
      },
      "source": [
        "print('Type X:', type(X))\n",
        "print('Type y:', type(y), end='\\n\\n')\n",
        "\n",
        "print('X Shape:', X.shape)\n",
        "print('y Shape:', y.shape)"
      ],
      "execution_count": 3,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Type X: <class 'numpy.ndarray'>\n",
            "Type y: <class 'numpy.ndarray'>\n",
            "\n",
            "X Shape: (17687, 224, 224)\n",
            "y Shape: (17687,)\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 318
        },
        "id": "1lB3uH71Xq9N",
        "outputId": "a6d317dc-6989-4d6c-e53e-08c86f32adab"
      },
      "source": [
        "index = 1113\n",
        "random_img = X[index]\n",
        "random_img_label = CATEGORIES[y[index]]\n",
        "\n",
        "print('Random Image from X:')\n",
        "plt.imshow(random_img, cmap='gray')\n",
        "plt.title(random_img_label)\n",
        "\n",
        "print()"
      ],
      "execution_count": 4,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Random Image from X:\n",
            "\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "1dO3PNSnE-nl",
        "outputId": "1c3cf8f6-b340-458a-d38d-43e5725f3bf6"
      },
      "source": [
        "# Reshape the size to reduce RAM consumption\n",
        "reshaped_size = 64\n",
        "print('Before Reshaping size:', X.shape)\n",
        "\n",
        "resized_X = []\n",
        "for img in X:\n",
        "  resized_X.append(cv2.resize(img, (reshaped_size, reshaped_size)))\n",
        "\n",
        "X = np.asarray(resized_X)\n",
        "del resized_X\n",
        "\n",
        "print('After Reshaping Size: ', X.shape)"
      ],
      "execution_count": 5,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Before Reshaping size: (17687, 224, 224)\n",
            "After Reshaping Size:  (17687, 64, 64)\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 301
        },
        "id": "8Q4o6WBzPCWf",
        "outputId": "bb389642-77b6-40b6-ff13-4b0f440e6a5a"
      },
      "source": [
        "plt.imshow(X[index], cmap='gray')\n",
        "plt.title(CATEGORIES[y[index]])\n",
        "print()"
      ],
      "execution_count": 6,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "31VBX6U-O-J4",
        "outputId": "de654dca-8b97-4f50-a4b8-3aab25e4f20b"
      },
      "source": [
        "# Reshape into 1d for Naive Bayes \n",
        "print('Before Reshaping to 1D:', X.shape)\n",
        "X = X.reshape(-1, reshaped_size*reshaped_size)\n",
        "print('After Reshaping to 1D: ', X.shape)"
      ],
      "execution_count": 7,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Before Reshaping to 1D: (17687, 64, 64)\n",
            "After Reshaping to 1D:  (17687, 4096)\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 318
        },
        "id": "bmUvzvmPTpiE",
        "outputId": "9ae00217-dc64-4ae7-a244-6903b5572c0d"
      },
      "source": [
        "same_img = np.reshape(X[index], (reshaped_size, reshaped_size))\n",
        "print('Double Check: Img still holds after reshape', same_img.shape)\n",
        "plt.imshow(same_img, cmap='gray')\n",
        "plt.title(CATEGORIES[y[index]])\n",
        "print()"
      ],
      "execution_count": 8,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Double Check: Img still holds after reshape (64, 64)\n",
            "\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "5LKT-1dvaRJh",
        "outputId": "4bae5c2e-e22d-4d51-bd8d-1f801694fa70"
      },
      "source": [
        "# Normalize the data\n",
        "X = X / 255.0\n",
        "print('Domain: [',X.min(),',', X.max(), ']')"
      ],
      "execution_count": 9,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Domain: [ 0.0 , 1.0 ]\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "P3MDlNG3ejzC"
      },
      "source": [
        "# Split dataset\n",
        "X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2, random_state=10)"
      ],
      "execution_count": 10,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "XAL4IC72NigB"
      },
      "source": [
        "#### GaussianNB w/o Hyperparameter tuning\n"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "MPYH5bLOE1IN",
        "outputId": "d8cfa0fb-d23c-4f6a-9627-c67422f092ed"
      },
      "source": [
        "# GaussianNB Building\n",
        "gaussianNB_model = sklearn.naive_bayes.GaussianNB()\n",
        "\n",
        "# Training\n",
        "gaussianNB_model.fit(X_train, y_train)"
      ],
      "execution_count": 11,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "GaussianNB(priors=None, var_smoothing=1e-09)"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 11
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 602
        },
        "id": "LQYyIKCMU-YJ",
        "outputId": "91652f74-1ffe-4369-86ae-2049d5598ae9"
      },
      "source": [
        "# Evaluation\n",
        "gnb_y_pred = gaussianNB_model.predict(X_test)\n",
        "\n",
        "print('GaussianNB Metrics', end='\\n\\n')\n",
        "accuracy = sklearn.metrics.accuracy_score(y_test, gnb_y_pred)\n",
        "print('Accuracy: %f' % accuracy, end='\\n\\n')\n",
        "\n",
        "recall = sklearn.metrics.recall_score(y_test, gnb_y_pred, average='micro')\n",
        "print('Recall: %f' % recall, end='\\n\\n')\n",
        "\n",
        "f1_score = sklearn.metrics.f1_score(y_test, gnb_y_pred, average='micro')\n",
        "print('F1 Score: %f' % f1_score)\n",
        "\n",
        "cm = sklearn.metrics.confusion_matrix(y_test, gnb_y_pred)\n",
        "# print('Confusion Matrix:')\n",
        "# print(cm, end='\\n\\n')\n",
        "plt.figure(figsize=(8,8), facecolor='w')\n",
        "sns.heatmap(cm, annot=True, fmt='d', linewidths=.5, square=True,cmap='Blues_r', yticklabels=CATEGORIES, xticklabels=CATEGORIES)\n",
        "plt.title('Confusion Matrix for GaussianNB')\n",
        "\n",
        "print()"
      ],
      "execution_count": 12,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "GaussianNB Metrics\n",
            "\n",
            "Accuracy: 0.779536\n",
            "\n",
            "Recall: 0.779536\n",
            "\n",
            "F1 Score: 0.779536\n",
            "\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAloAAAHBCAYAAACxEhX3AAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4yLjIsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+WH4yJAAAgAElEQVR4nOzde3zP9f//8dsODGMY2yznQy3KYcymnGrEMMNUpM8SfZyPFZEKyaGhFCrHogPKYbY5f53lsJlDQiLJ2WYWs7HNtvfvj/fP+9Maw7xezXS/dnlfLt7P5+vwfL/Xtof78/l+srNYLBZERERExHD2eT0AERERkYeVCi0RERERk6jQEhERETGJCi0RERERk6jQEhERETGJCi0RERERkzjm9QBERETk4ZOSbs51C+WzykWJloiIiIhJVGiJiIiI4SwWcx53KzQ0FH9/f7y8vDh69Kit/cSJE3Tq1ImWLVvSqVMn/vjjj/vuy4kKLREREXnoNGvWjO+++46yZctmaR81ahRdunRh7dq1dOnShZEjR953X05UaImIiIjhLCb9d7d8fHzw9PTM0nbp0iUOHz5MYGAgAIGBgRw+fJiEhIRc991JPltSJiIiIv9miYmJJCYmZmt3cXHBxcUlx3PPnz+Ph4cHDg4OADg4OODu7s758+exWCy56nN1dc3xniq0RERExHj3sJ7qXsyfP5/p06dna+/fvz8DBgww56b3QYWWiIiIGM6kOouuXbvSoUOHbO13SrMAPD09iY2NJSMjAwcHBzIyMoiLi8PT0xOLxZKrvjvRGi0RERHJN1xcXChXrly2x90UWqVKlaJ69eqsWLECgBUrVlC9enVcXV1z3XcndhbLvXxYUkREROTOklLNKS+KOtnd1XFjx45l3bp1xMfHU7JkSUqUKMHKlSs5fvw4w4cPJzExERcXF0JDQ6lSpQpArvtyokJLREREDJfXhdaDQmu0RERExHD3shXDvVGhJSIiIv92mi8DtBheRERExDRKtERERMRwCrSslGiJiIiImESJloiIiBhOexpYqdASERERw5n3qcP8RVOHIiIiIiZRoiUiIiKG09ShlRItEREREZOo0BIRERExiQotEREREZOo0BIRm5SUFHr37k29evUYOHBgrq8TERFB9+7dDRxZ3vjvf/9LWFhYrs6dMmUKfn5+NGzY0OBRPTjatGlDVFRUXg9DHlAWizmP/EaFlkg+FBkZSXBwMN7e3jRq1Ij//ve/xMTE3Pd116xZQ3x8PFFRUUydOjXX1wkKCuLLL7+87/H8XVRUFF5eXvTr1y9L+5EjR/Dy8iIkJOSurjNt2jSGDBlyx+PmzJlDhw4d7nmc586d46uvvmLVqlVs3779ns+/FYvFwrfffkvbtm2pXbs2DRs2JCQkhJUrVxpy/dxYuXIlfn5+932dZcuW4eXlxezZs7O0N2nSxFbITZs2jSeeeAJvb2+8vb1p1aoVa9euve97i3ksJv2X36jQEslnvvrqK8aPH0/v3r3Zvn07mzZtokuXLmzYsOG+r33u3DkqVaqEo+OD+4FkV1dX9u/fz59//mlrCwsLo1KlSobdw2KxkJmZmevzz507R4kSJShVqtQ9n5uenn7L9rFjxzJ//nyGDx9OVFQUW7duZdCgQWzbti3X43yQlChRgjlz5pCUlHTbY1q1asW+ffvYt28fI0aMYOjQocTHx/+DoxS5dyq0RPKRq1evMnXqVEaOHEmLFi0oUqQIBQoUwN/fn2HDhgGQlpbGuHHjaNSoEY0aNWLcuHGkpaUB1kSoSZMmfPnllzz11FM0atSIpUuXAjB16lQ+//xzVq9ejbe3N4sXL86W/Jw5cwYvLy9bMbBs2TKaNWuGt7c3/v7+RERE2Npfeukl23l79+6lY8eO1KtXj44dO7J3715bX0hICJ988gmdO3fG29ub7t27k5CQcNv3oECBAjRr1oxVq1YBkJGRwapVq2jbtm2W48aOHUvTpk2pW7cuwcHBtsRv69atzJw50/Y6g4KCbOOYMmUKnTt3pnbt2pw+fZqQkBAWL14MwKhRoxgwYIDt+pMmTaJr165Y/jaXsWPHDrp3705cXBze3t4MHz4cgA0bNtCmTRt8fHwICQnh+PHjtnP8/f2ZNWsWbdu2pU6dOtmKrRMnTrBgwQI+/vhjGjZsSKFChXBwcMDHx4cPP/zQdtzSpUtp1aoV3t7eNGvWjEWLFtn6/v41AfDy8uLkyZMAbNmyhdatW+Pt7U3jxo2ZO3cuAAkJCfTq1QsfHx98fX3p0qWLrQj19/dnx44dABw4cIBOnTrh4+NDo0aNGDNmjO3/u5v3WrhwIS1atMDHx4f3338/y3tXpUoVvL29mTdv3t+/5LfUuHFjnJ2dOXXq1F0dL/88TR1aqdASyUf27dtHamoqzz333G2P+eKLL/jpp58IDw8nIiKCn3/+mc8//9zWHx8fz9WrV9m6dSvjxo1jzJgxXLlyhYEDB9KrVy9bavDCCy/kOJZr164xduxYZs+ezb59+1i0aBHVq1fPdtzly5fp1asXISEhREVF0a1bN3r16pUlkVqxYgUTJkxg586d3Lhx447Tju3bt2f58uUA/Pjjjzz22GN4eHhkOaZmzZosX76c6OhoAgMDGTRoEKmpqTRp0iTL67xZHAKEh4fzwQcfsHfvXh555JEs1xs+fDhHjx5l2bJlxMTEsGTJEkJDQ7Gzs8ty3NNPP83s2bNxd3dn3759fPjhh5w4cYI333yTESNGsHPnTpo0aULv3r2zFCIrV65k1qxZxMTEZEsUd+3ahaenJzVr1szxfSlVqhQzZ85k7969TJgwgQkTJnDo0KEcz7npnXfeYcyYMezbt48VK1bQoEEDwJqgenh4sHPnTrZv384bb7yR7TUD2Nvb8/bbb7Nr1y4WLVrEzp07WbBgQZZjNm/ezJIlS4iIiGD16tXZ0rhBgwYxf/58Ll++nONYLRYLmzdv5saNG1SrVu2uXp9IXlGhJZKPXL58mZIlS+Y4tRcZGUm/fv0oVaoUrq6u9OvXL0sx4ejoSL9+/ShQoABNmzalSJEinDhxIlfjsbe359ixY6SkpODu7s6jjz6a7ZjNmzdTsWJF2rdvj6OjI4GBgVSpUoVNmzbZjgkODqZy5coUKlSIgIAAfvnllxzvW7duXa5cucLvv//O8uXLadeuXbZj2rVrZ3uvunfvTlpa2h1fZ4cOHXj00UdxdHSkQIECWfoKFy7MxIkT+fDDDxk6dCjvvfceZcqUyfF6N61atYqmTZvSsGFDChQowGuvvUZKSgr79u2zHRMSEoKnpyeFChXKdv6ff/5J6dKls7Q1adIEHx8fatasydmzZwF45plnqFChAnZ2dvj6+tKwYcO7Xrvn6OjIb7/9RlJSEsWLF+eJJ56wtV+8eJFz585RoEABfHx8blloPfnkk9SpUwdHR0fKlStHp06d2L17d5ZjevTogYuLC4888gh+fn4cOXIkS3/16tVtheqtrFmzBh8fH7y9venTpw+9evXCxcXlrl6f/PMsJj3yGxVaIvlIiRIl+PPPP2+7jgcgLi4uSxrzyCOPEBcXl+Uafy3UChcuzLVr1+55LEWKFGHKlCksWrSIRo0a0bNnzyzTYbcbz80xxcbG2p67ubnd83iCgoL47rvviIqKumXCN3fuXFq1akW9evXw8fHh6tWrWVK0W/H09Myxv3bt2pQrVw6LxUKrVq3uOMab/v4e2Nvb4+npmeU9yOneJUqU4OLFi1natm7dyq5du0hLS7NNwW3ZsoUXX3wRX19ffHx82Lp16x1f801Tp05ly5YtPPvss/znP/+xFYGvvfYaFStWpHv37jRr1oxZs2bd8vwTJ07Qq1cvGjZsSN26dZkyZUq2e//965ycnJztOgMHDmThwoW3XHsVEBBATEwM+/fv5//+7/8IDw/PMj0qDxhVWoAKLZF8xdvbm4IFC7J+/frbHuPu7s65c+dsz8+fP4+7u3uu7le4cGFSUlJsz//+y69x48Z89dVX/Pjjj1SpUoX33nvvjuO5Oaa/T/Xdq3bt2rFgwQKaNm1K4cKFs/TFxMQwZ84cPvnkE3bv3k1MTAzFihWzFSS3SmRyar/pu+++48aNG7i7uzNnzpy7Huvf3wOLxZLtPcjp3g0aNODChQv8/PPPtz0mLS2NgQMH0r17d7Zv305MTAxNmjSxvea/fy3/XrjVqlWLL774gh07dtC8eXMGDx4MQNGiRRk+fDgbNmzgiy++4KuvvmLnzp3Z7j969GiqVKnC2rVr2bt3L6+//nq29Wt3o2rVqrRo0YIZM2bkeFy5cuVo3LhxlmRU5EGkQkskHylWrBgDBw5kzJgxrF+/nuvXr3Pjxg22bNnCxIkTAeveRl988QUJCQkkJCTw2WefZVsofreqV6/O7t27OXfuHFevXmXmzJm2vvj4eNavX8+1a9coWLAgRYoUwd4++4+Upk2b8scffxAZGUl6ejqrVq3it99+45lnnsnVmG4qX74833zzja0g+Kvk5GQcHBxwdXUlPT2d6dOnZ/k0W6lSpTh79uw9fbLwxIkTfPLJJ0yaNImJEycyZ86cO05x3tSqVSu2bNmSZQ1awYIF8fb2vqvzq1SpQqdOnXjjjTfYvn07KSkpZGRkZJl6TEtLIy0tDVdXVxwdHdmyZUuWrSUef/xxjh07xi+//EJqairTpk3Lcm5ERARXr16lQIECODs7276WmzZt4uTJk1gsFooVK4aDg8Mti8Lk5GScnZ1xdnbm+PHjLFy48K5e263069ePpUuXcvXq1dsec+HCBX788Uet0XqAaXsHqwf3M9wickvdu3endOnSfP755wwZMgRnZ2eeeOIJevfuDUDfvn1JTk62fZouICCAvn375upeDRs2pHXr1gQFBVGyZEl69OjBxo0bAcjMzGTevHkMGzYMOzs7qlevzujRo7Ndo2TJksyYMYPx48czevRoKlasyIwZM3B1dc3dG/AXPj4+t2xv1KgRjRs3pmXLlhQpUoSuXbtmmZoLCAggIiICPz8/ypUrd8dNSdPT0xk6dCg9evTg8ccfB+D111/nrbfeYunSpRQsWDDH86tUqcKkSZP44IMPiI2NpXr16syYMeOO5/3VqFGj+Oabb/jwww85efIkLi4uVKpUiSlTpvDII49gb2/Pu+++y+DBg0lLS+PZZ5/F39/fdn7lypXp168fr776KoUKFeKNN97g+++/t/Xf/CBARkYGlStXZtKkSQCcPHmSDz74gISEBFxcXHjppZdsC+X/atiwYbz33nvMnTuX6tWr07p1a3bt2nXXr++vypcvT7t27bIVa6tXr7ZtY+Ls7Iy/vz/9+/fP1T1E/il2ltxkuyIiIiI5iE28Ycp1PVwK3PmgB4gSLRERETGcUhwrrdESERERMYkSLRERETGeIi1AiZaIiIiIaZRoiYiIiOHy41YMZlChJSIiIobTngZWKrTkH3fBpI/8St4q41KAwt7a0+hhdH3fdMr1XZ7XwxATnPm8fV4P4aGnQktEREQMp0DLSovhRUREREyiREtEREQMpzVaVkq0REREREyiREtERERMoEgLVGiJiIiICTR1aKWpQxERERGTKNESERERwynQslKiJSIiImISJVoiIiJiOK3RslKhJSIiIobTPyptpalDEREREZMo0RIRERHjKdAClGiJiIiImEaJloiIiBhOgZaVCi0RERExnD51aKWpQxERERGTKNESERERw2l7ByslWiIiIiImUaIlIiIixlOgBSjREhERETGNEi0RERExnAItKxVaIiIiYjht72ClqUMRERERkyjREhEREcNpewcrJVoiIiIiJlGiJSIiIsZToAWo0BIRERETqM6y0tShiIiIiEmUaImIiIjhtL2DlRItEREREZMo0RIRERHDaXsHKxVaIiIiYjzVWYCmDkVERERMo0RLREREDKdAy0qJloiIiIhJlGiJiIiI4bS9g5USLRERERGTKNESERERw2l7BysVWiIiImI81VmApg5FRERETKNES0RERAynQMtKiZaIiIiISZRoiYiIiOG0vYOVCi0RERExnD51aKWpQxERERGTKNESERER4ynQApRoiYiIiJhGhZaIiIgYzmLS415s2rSJ9u3b065dO4KCgli3bh0AJ06coFOnTrRs2ZJOnTrxxx9/2M7JqS83VGiJiIiI4SwWcx53f38Lb731FhMnTiQ8PJyJEycybNgwMjMzGTVqFF26dGHt2rV06dKFkSNH2s7LqS83VGiJiIjIQ8ne3p6rV68CcPXqVdzd3fnzzz85fPgwgYGBAAQGBnL48GESEhK4dOnSbftyS4vhRURExHBmbe+QmJhIYmJitnYXFxdcXFxsz+3s7Pjkk0/o27cvRYoUITk5mVmzZnH+/Hk8PDxwcHAAwMHBAXd3d86fP4/FYrltn6ura67Gq0JLRERE8o358+czffr0bO39+/dnwIABtufp6enMnDmTzz//nHr16rFnzx4GDx7MxIkT/8nhqtASuR9LFn7DiuVLsVgsBLZ/nhe6hDD3i2n8uHUj9nb2lHB15e1R4yjt5m4755dDP9Pvtf8wctwknmnWIg9HL3drxqiXadXkSS4mXMXnhfG29j6dm9LrxcZkZFpYs+0g73wanoejlLvhWbIwn3atS+liTlgssGD7H8zd9DtDAqvTsnYZMjMhPimVN77eS+yVFHo3r0aH+uUBcHCw49Eyxaj91iouX7uRx68kHzBpe4euXbvSoUOHbO1/TbMAfvnlF+Li4qhXrx4A9erVo3Dhwjg5OREbG0tGRgYODg5kZGQQFxeHp6cnFovltn25pTVadzBlyhRGjRple75p0ya8vLw4duyYra1Xr14sXryYDRs2EBoaCsCZM2f4/vvvs1zL39+fo0ePGja2W93jr6KiovDy8rKN6aaQkBC8vLxITk7O1X2XLVvGwIEDc3Xuw+T3346xYvlSZsxfyNwFS9n54xbOnD5F55BufLUwjLkLlvJUo6bMn/OF7ZyMjAxmTp+Cj9/TeThyuVffRO6iXb/PsrQ18XmUwGdq4tvpQ+o9P45Pvt6QR6OTe5GRkcmYpQfx/2AjQZO20rVJFR4tU4wZ64/x3LhNtJywiQ0/X2Bway8AZqz/jZYTrO0fhh9m17F4FVl5zMXFhXLlymV7/L3QKlOmDBcuXOD3338H4Pjx41y6dImKFStSvXp1VqxYAcCKFSuoXr06rq6ulCpV6rZ9uaVC6w4aNGhAdHS07Xl0dDS1a9e2tWVkZLBnzx78/Pxo1qwZw4YNA+Ds2bM5FkFGuJt7VK5cmQ0bNpCRkQHA6dOnuXbtmqnj+rc4+cfvVH+yJoUKFcbR0ZHadX3Yumk9zkWL2o5JuX4d7Oxsz5d9v4Cmzz5HyZK5/6aVf972vcdJuJL1+6bnC42Z/NX/kXYjHYCLfyblxdDkHsUlpnLw9BUAklPTOXbhKmVKFCIpJd12TGEnh1t+uq29T1nCY878U0PN9/J6ewc3NzdGjx7NoEGDCAoK4vXXX2f8+PGUKFGC0aNH8+2339KyZUu+/fZb3n//fdt5OfXlhgqtO/D29ubMmTPEx8cDsHv3bvr27UtUVBQAhw8fpmjRolSoUCFL0jNmzBiOHz9Ou3btsqQ/q1evplOnTvj7+/Ptt9/a2g8cOECnTp1o27YtnTp14sCBA4A1lQoODrYd99fnt7vHXxUpUgRvb29+/PFHAMLCwmjfvn2WY0JDQ+nYsSNBQUF07dqVs2fPAnDp0iVeffVV2rZtS9u2bRk/fny2658/f57g4GBWrVp1D+/qw6Fy1Woc2L+XK5cvk5JynV07thEXewGA2Z9/yvNtmrF+zUpe69UfgItxsWzbvIF2z3fKy2GLQapVdKehd1W2fj2EdXMGUa9Ghbwektyjcq5FeLJ8cfb98ScAbwVVJ3pcCzrUL8/kFb9kObZQAQeeqeHBqn3n8mKo+VJeb+8AEBQURGRkJBEREURERNC8eXMAqlatyuLFi1m7di2LFy+mSpUqtnNy6ssNFVp3UKhQIWrVqkV0dDRJSUlcv36dxo0bc+TIEcCacPn6+mY7b+TIkVStWpXw8HCmTp1qa09JSeH777/n66+/5qOPPiI5OZm0tDQGDhzI4MGDiYyMZNCgQQwcOJC0tLQcx3a7e/xdhw4dCAsLw2KxsHLlStvHVm/q0aMHS5cuJSIigsDAQCZPngxAZGQkFSpUIDIyksjISPr165flvCNHjtCzZ09GjBhB69atc34jH0KVKlelyyvdGTKgJ0MH9qbaY1442Fu/pXr0HcSSlRtoHtCGZT8sAGDax6H0GvA69vb6tnsYODrY41rcmSavTGbElOV8O7F7Xg9J7kERJwdm9fRl9JKfbWnWxIhf8H1nHWG7T9OtadZfrs/VKsPu3xM0bSj3TIvh74Kvry9RUVE4OztTr149HBwcqFixIseOHSM6OpoWLe5+QfPNguTmfPKFCxdIT0+nQIECPPXUUwA8/fTTFChQgBMnThgyfj8/P95//33Wr1/PY489RsmSJbP0b926lQULFnDt2jXS0/8Xn9euXZt58+YRGhqKr68vjRo1svX9+uuv9O/fn5kzZ1K1alVDxpkftWnXkTbtOgIw67NPcHMvk6X/uVaBDBvUh+69+vPrL4cY885QAK5c/pNdO7bh4OBA42ea/ePjlvt3NvYyyzfsByDm0EkyMy2ULlmUeE0hPvAc7e2Y1cOXsOjTrN5/Plt/WPQZvu73FB+tPGJra1evLOG7NW14L8za3iG/0V+t74Kfnx/R0dHs3r2b+vXrA1C/fn127tzJnj17bplo3Y6Tk5Ptzzc/0ZATBwcHLH/JSlNTU+9x9Na9RFq1asW7776b7ZMaZ8+eZcKECXz00UesWLGC8ePH25I0b29vwsLCePLJJwkPD+eVV16xnefh4UHx4sVtU6j/Vn8mXAIg9sJ5tm3aQPOA1pw5ddLW/+OWjVSoVBmA78PX8n3EOr6PWEdT/xa8PuxdFVn5WOTmAzSt/xgA1Sq4U7CAo4qsfGJyiDe/XUhi9sbjtrbKbs62P7es7cnxC1dtz4sVcqTBo6VZeyB7USZyJ0q07oK3tzdnz55l3bp1hISEAODj48Pbb7+Ni4sL5cuXz3ZO0aJFSUq6ux+6lStX5saNG+zatYsGDRqwc+dO0tPTqVy5Mn/++SenT5/mypUruLi4sHLlylzdo1OnThQuXJgmTZpkaU9KSqJAgQK4ubmRmZnJokWLbH2nT5+mTJkytGnTBh8fH5577jkyMzMBKFGiBJ999hk9evQgNTWVbt263dU4HjbvDXudxCuXcXR0ZPBb71CsmAsTPxjJ6ZN/YGdvh0eZR3jz7fv75xsk782f8CqN6z1K6RJF+W3NB3wwYxXzl+9k5uiXiVk8grQbGfx35Dd5PUy5C/WruvK8XwV+OXuFtW8/C0BoxGE6P12RKh5FsVgsnEm4ztsL9tvOCajzCFt+ieN6Ws5/MZa/UaAFqNC6K05OTtSuXZvY2Fg8PDwAqFmzJrGxsQQEBNzyHC8vLypXrkxgYCBVqlTJcQ1VwYIFmTp1KuPGjePatWsUKVKETz/9lIIFC+Lh4UG3bt0IDg6mdOnS1K9f37a1xL3cw8PDgx49etxynAEBAbRu3ZqSJUvStGlTYmJiAOv6s3nz5mFvb09mZibvv/9+lvVFxYoVY+7cufTu3Ztr165lW8P1bzB99tfZ2j6Y+Mkdz3t79DgzhiMm6fr2vFu2d383+9dfHmy7jydQru/ybO0bD8Xe9pzFu06xeNcpM4f1UFKdZWVnsdzrGn6R+3MhUYtJH0ZlXApQ2Lt/Xg9DTHB93/RbFieS/535vP2dD8ql6N+vmHJd3yrFTbmuWZRoiYiIiOEU41hpMbyIiIiISZRoiYiIiOG0vYOVCi0RERExnuosQFOHIiIiIqZRoiUiIiKGU6BlpURLRERExCRKtERERMRwmdrfAVCiJSIiImIaJVoiIiJiOOVZViq0RERExHCaObTS1KGIiIiISZRoiYiIiOG0M7yVEi0RERERkyjREhEREcNlKtACVGiJiIiICTR1aKWpQxERERGTKNESERERw2l7ByslWiIiIiImUaIlIiIihtMaLSsVWiIiImI4ferQSlOHIiIiIiZRoiUiIiKG09ShlRItEREREZMo0RIRERHDaXsHKyVaIiIiIiZRoiUiIiKGU6JlpUJLREREDJepxfCApg5FRERETKNES0RERAynqUMrJVoiIiIiJlGiJSIiIobThqVWKrRERETEcJo6tNLUoYiIiIhJlGiJiIiI4bS9g5USLRERERGTKNESERERw2mNlpUKLRERETGc6iwrTR2KiIiImESJloiIiBjOorlDQImWiIiIiGmUaImIiIjhMvN6AA8IJVoiIiIiJlGiJSIiIobTGi0rFVoiIiJiOJVZVpo6FBERETGJEi0RERExnKYOrZRoiYiIiJhEiZaIiIgYTts7WKnQEhEREcNp6tBKU4ciIiIiJlGiJSIiIoZToGWlREtERETEJHYWTaKKiIiIwb7dc8aU6/6nXjlTrmsWTR2KiIiI4TKV4wAqtCQPFPYbmtdDEBNcj5pESnpej0LMUMgRCgd9kdfDEBNcj+iT10N46KnQEhEREcMpz7LSYngRERERkyjREhEREcPps3ZWKrRERETEcPoneKw0dSgiIiJiEiVaIiIiYjjNHFop0RIRERExiRItERERMZw2LLVSoiUiIiJiEiVaIiIiYjgFWlYqtERERMRwmjq00tShiIiIiEmUaImIiIjhMhVoAUq0REREREyjREtEREQMpyVaViq0RERExHCZqNICTR2KiIiImEaJloiIiBjuQZg6TE1NZfz48ezcuRMnJyfq1KnDBx98wIkTJxg+fDiXL1+mRIkShIaGUqlSJYAc+3JDiZaIiIg8lCZNmoSTkxNr164lMjKSQYMGATBq1Ci6dOnC2rVr6dKlCyNHjrSdk1NfbqjQEhEREcNlWsx5JCYmcubMmWyPxMTELPdPTk5m+fLlDBo0CDs7OwBKly7NpUuXOHz4MIGBgQAEBgZy+PBhEhIScuzLLU0dioiIiOHM2hl+/vz5TJ8+PVt7/2M/u9UAACAASURBVP79GTBggO356dOnKVGiBNOnTycqKgpnZ2cGDRpEoUKF8PDwwMHBAQAHBwfc3d05f/48Fovltn2urq65Gq8KLREREck3unbtSocOHbK1u7i4ZHmekZHB6dOnqVGjBsOGDeOnn36id+/efPrpp//UUAEVWiIiImICsxbDu7i4ZCuqbsXT0xNHR0fbNGDt2rUpWbIkhQoVIjY2loyMDBwcHMjIyCAuLg5PT08sFstt+3JLa7RERETkoePq6oqfnx/bt28HrJ8mvHTpEpUqVaJ69eqsWLECgBUrVlC9enVcXV0pVarUbftyy85ieRA+gCn/JoX9hub1EMQE16MmkZKe16MQMxRyhMJBX+T1MMQE1yP6mHbtSZt/N+W6Q5+pctfHnj59mhEjRnD58mUcHR0ZPHgwTZs25fjx4wwfPpzExERcXFwIDQ2lShXrdXPqyw1NHYqIiMhDqXz58nzzzTfZ2qtWrcrixYtveU5OfbmhQktEREQMpwkzKxVaIiIiYrhM1VmAFsOLiIiImEaJloiIiBhOiZaVEi0RERERkyjREhEREcNZUKQFKrRERETEBJo6tNLUoYiIiIhJlGiJiIiI4bSNlpUSLRERERGTKNESERERw2Uq0gJUaImIiIgJtBjeSlOHIiIiIiZRoiUiIiKG08yhlRItEREREZMo0RIRERHDaTG8lRItEREREZMo0RIRERHDKdCyUqElIiIihsvM6wE8IDR1KCIiImISJVoiIiJiOC2Gt1KiJSIiImISJVoiIiJiOAVaViq0RERExHD6tw6tNHUoIiIiYhIlWiIiImI4i+YOASVaIiIiIqZRoiViIHt7O7bPG8S5i1fo+OZXfPX+S9StXo4b6ZnEHD5F/wlLSc/QNn4PqpHvvs3WLZtxdS3FsvAVAFy5fJm3hrzOubNneaRsWSZ99AkuxYtz4vfjjHx3BL8cPsSAQa/TtdtrWa6VkZHBSy92xN3Dg+mfz8yLlyO3MWPgM7TyqcTFK9fxGfA9AONffYrWvhVJS8/kxPkr9Jy6iSvJaXRu+iiDO9SxnVuzUimeen0xB05cyqvh5xtao2V1x0TL39+fo0eP/hNjMURiYiKzZ8/O0hYSEsKmTZvu+honT57E29ub9PR0wBp/PvXUU4SGhtqOWbNmDV26dDFm0EBsbCwhISGGXQ9g2bJleHl58d1339naLBYLzZo1w8/PL9fXnTZtWpb3Qv6nf6fG/PpHnO35orX7qP3iJHy6fERhpwJ0a5f7913M1659MF/MnJOl7cs5s/D1e4rI1evw9XuKuXNmAeBSvATD3n4nW4F103fffE2VKlVNH7Pcu282/Eq70SuytG3Yf5p6/b/Hd+APHDt3haHP1wVg0ZZjNBi8mAaDF/PalA38EZuoIusuZVrMeeQ3D8zU4c2i5qbMzMxcze8mJiYyZ86cOx+Yg4oVK+Li4sKhQ4cA+O233yhbtiwxMTG2Y6Kjo/H19b2v+9yUnp6Oh4cH33zzjSHX+6saNWqwfPly2/OoqCiKFy9u+H0EyroXJ6Dh43wVHmVrW7vjiO3PMYdOU9Zd7/2DrJ5PfVz+9v2xadMGgtq3ByCofXs2bVwPQKlSpXiyZi0cHbNPDMReuMC2rZvp0PF58wct92z7ofMkJKVmaduw/wwZ//+3ePSvsZQt5ZztvBebPMribb/9I2OUh8c9FVohISGEhoby0ksv0axZMyZPnmzri42NZcCAAbRt25a2bdsyc6Y1Ko+Pj6dfv3629r/+0vf392fy5Mk8//zzjBw5kmnTpjFw4EC6d+9O69atSUxMZMuWLXTu3Jng4GA6derE/v37becvWbKEoKAggoKC6NixI/Hx8YwZM4arV6/Srl07OnfunGX8sbGxNGrUiNTU/32D9e7dm8jIyGyv1dfXl6go6y/M6OhoWrRowY0bN0hKSrK13UyFZs2axfPPP0+HDh3o3bs3Fy9eBGDnzp106tSJ9u3b07ZtW1auXJnlvRw3bhwvvvgiffr04cyZM1lSJi8vL2bMmEHHjh1p1qwZa9eutfWtXbuWgIAA2rdvz4wZM/Dy8iI5OfmWX7Py5ctTqFAhfvvN+sMhLCyMDh06ZDnmzTffJDg4mLZt29KvXz+uXLkCwO+//06nTp0ICgoiMDCQuXPnZrv+r7/+Stu2bYmOjr7l/f9NJr0exDvTV95yN2RHB3tealWX/9v1ax6MTO5HwqVLuLm5A1C6tBsJl+6cZkz8cDyvvzkUe/sH5u+ycg9eaf44a/eeytb+fKOq/LBVhdbdslgspjzym3teo3X+/Hm+++47kpOTad68Oc8//zyVKlViyJAhNG3alGnTpgGQkJAAwNixY3n00Uf57LPPiIuLIzg4mBo1avDYY48BkJSUxJIlSwDrlNSBAwdYtmwZrq6unDp1is8//5y5c+dStGhRjh07Ro8ePdi8eTNRUVHMnDmTBQsW4ObmRnJyMo6OjowcOZKOHTsSHh6ebeweHh7Ur1+fVatW0aFDB86cOcPBgweZOnVqtmP9/PxYs2YNPXv2JDo6mldffZXz588TExNDrVq1OHXqFN7e3oSHh3P69Gl++OEH7O3tWbBgAR9++CEfffQRNWrUYMGCBTg4OBAfH09wcDCNGjWyJUqnT59mwYIFODo6cubMmWxjKFq0KEuXLmXPnj0MHjyYli1bEh8fz8iRI/n++++pVKkS8+bNu+PXrH379oSFhdG3b1/27NlD3759mT59uq3/nXfewdXVFYApU6Ywe/ZshgwZwoIFC/D396dXr14AtgLsph07djBhwgSmTJlCtWrV7jiOh1mrhtWJS0hi35GzNK5bJVv/p28Fs33/CbbvP5EHoxOj2NnZgZ1djsds2bwJV1dXajzxJLujo3I8Vh48b71Ql4yMTBZtPpalvf5j7lxLTefwqYQ8GpnkV/dcaAUEBGBvb0+xYsWoWrUqp06dws3NjX379vHVV1/Zjrv5i3vnzp0MHz4cAHd3d5o2bUpUVJSt0Gr//yP5m5o0aWI7d9u2bZw6dYqXX37Z1p+enk58fDybN2+mXbt2uLm5AeDsnD3mvZWQkBAmTJhAhw4dWLRoER07dqRgwYLZjvPz82P8+PGkp6dz6NAhatasyfnz54mKiiIlJYVatWpRqFAhNm7cyMGDB20pUUZGBkWLFgWsxeaIESM4efIkDg4OXLlyhRMnTlCnjnVhZdu2bW857XBT69atAahTpw5xcXGkpqby008/UaNGDSpVqgRAx44dmTBhQo6vOSAggODgYCpVqkTz5s1xcHDI0h8eHk5kZCQ3btzg2rVrtmvXr1+fSZMmcf36dfz8/GjQoIHtnB9//JFt27Yxd+5cPDw8crz/v8FTtSsR2KQGAU8/jpNTAVycnfhy9Et0H72QEa89h1tJZzoNW5rXw5RccC1ViosX43Bzc+fixTjbz6fb2b9vL5s3b+THbVtJTU0lOTmJt4cNYULo5BzPk7z3H38vWtevSKt3s89yvNC4Gj9o2vCe6GM/VvdcaDk5Odn+7ODgQEZGxn0NoEiRIlme/71gaty4MRMnTryve/xV3bp1ycjIYM+ePYSFhdnStL8rX748xYsXJzIykgoVKuDo6Ej9+vX58ssvSU1Nta3Pslgs9OnTh+efz74WY/To0fj7+zN9+nTs7Oxo2bJllmnLv7/2v7v5Xt8sjP6+ju1uOTs7U6dOHSZPnszXX3+dpS8mJoaFCxeyaNEiXF1diYyM5IcffgCgZcuW1KlTh+3btzN79myWLl1qmy6uXLkyx44d4+DBgyq0gJGfr2bk56sBaFy3CoNfbkr30Qt5NciX5xo8Rqv+M/Nl5C3wzLP+RCxfzms9ehKxfDnPPtssx+MHvf4mg15/E4Dd0VHMn/eliqx84Lm65XkjuA4tRoRzPS3rz1o7O+jYqCrNhi+/zdkit2fIAgJnZ2e8vb2zTGPdnDp86qmnbL+4L168yJYtW7IkIzlp2LAh27Zt49ix/0W4Bw4cAOCZZ54hPDyc+Ph4AJKTk0lNTaVo0aKkpKTkWJSEhITwxhtv4O3tjaen522P8/X1ZcaMGbaiys3NjWvXrrFlyxbbeip/f38WLFhgm1ZLS0vjyBHrAuirV69StmxZ7Ozs2L59OydPnryr152T2rVrc/jwYU6dsq4fCAsLu6vzevTowYABA/Dy8srSnpiYSNGiRSlRogRpaWksXfq/1OXkyZO4ubkRHBxMv379+Pnnn219ZcuW5csvv+Tjjz9m1apV9/26HlbThgXj7lqMzXMGsOub13n7teZ5PSTJwbAhb/BKl86c/OMEz/k3YdnSxXT/b0927dxO21YtiNq1g+7/7QlA/MWLPOffhG/mf8XsmV/wnH8T2xpOebDNH9KczRM78FjZ4vz2ZQhdn3ucKb0aU6xwQVaMacuuT15gap8mtuMbPfEIZ+KT+SP2ah6OOv/RGi0rw/bRmjx5Mu+//z6BgYHY29sTGBhIz549effddxk5ciRt27YFYMiQITz66KN3dc1KlSoxadIk3nnnHVJSUrhx4wZ169alVq1a+Pn50bNnT7p164adnR0FCxZkxowZlC5d2rbwvnjx4ixatCjbddu0acOYMWPuuD2Dn58fy5cvp379+ra2unXrEh4ebpv+a9++PZcvX+Y///kPYP0f66WXXuLxxx/nzTff5P3332fatGnUrFkzW5GTG6VLl2b06NH06NGDwoUL88wzz1CgQAEKFy6c43nVqlW75Tqqxo0bExERQcuWLSlZsiQ+Pj62gmr16tVERkZSoEAB7OzsGDFiRJZzPT09mTdvHq+99hopKSkEBwff9+t7GGzb+zvb9v4OQLGGw/N4NHIvQid/fMv22V/Oz9ZW2s2N/9u4Ncfr1ff1o76vtvR40HSdvD5b2/z/O3KLI622HTxH06HLzBzSQykf1kSmsLPkx/LwPsXExDB69GgiIyOti1vzmaSkJNs6sKVLl7JkyRIWLlyYx6O6e4X9hub1EMQE16MmkZK72W15wBVyhMJBX+T1MMQE1yP6mHbtPksPm3LdLzrWMOW6ZvnX7Qw/YsQIduzYQWhoaL4ssgC++eYb1qxZQ0ZGBsWLF2fs2LF5PSQREZEsbrXVzb/Rv67QGj9+fF4P4b716dOHPn3M+1uIiIiIGONfV2iJiIiI+RRoWanQEhEREcP9C5eA35L+fQgRERERkyjREhEREcMp0LJSoiUiIiJiEiVaIiIiYjht72ClQktEREQMpzLLSlOHIiIiIiZRoiUiIiKG0/YOVkq0REREREyiREtEREQMl6lAC1CiJSIiImIaJVoiIiJiOK3RslKhJSIiIoZTnWWlqUMRERERkyjREhEREcNp6tBKiZaIiIiISZRoiYiIiOG0vYOVCi0RERExnKYOrTR1KCIiImISJVoiIiJiOOVZVkq0REREREyiREtEREQMl6k1WoAKLRERETGB6iwrTR2KiIiImESJloiIiBhO2ztYKdESERERMYkSLRERETGcAi0rJVoiIiIiJlGiJSIiIobT9g5WKrRERETEcKqzrDR1KCIiImISJVoiIiJiOG3vYKVES0RERMQkSrRERETEcJkKtAAVWiIiImICC6q0QFOHIiIiIqZRoiUiIiKG01p4KyVaIiIiIiZRoSUiIiKGs1gspjxyY/r06Xh5eXH06FEA9u/fT1BQEC1btqR79+5cunTJdmxOfbmhQktEREQMl2kx53GvDh06xP79+ylbtqx1XJmZDB06lJEjR7J27Vp8fHyYPHnyHftyS4WWiIiI5BuJiYmcOXMm2yMxMTHbsWlpaYwZM4bRo0fb2g4ePIiTkxM+Pj4AdO7cmTVr1tyxL7e0GF5EREQMZ9bO8PPnz2f69OnZ2vv378+AAQOytH366acEBQVRrlw5W9v58+d55JFHbM9dXV3JzMzk8uXLOfaVKFEiV+NVoSUiIiL5RteuXenQoUO2dhcXlyzP9+3bx8GDBxkyZMg/NbRbUqElIiIihjNrewcXF5dsRdWt7N69m+PHj9OsWTMALly4wGuvvUZISAjnzp2zHZeQkIC9vT0lSpTA09Pztn25pTVaIiIi8tDp2bMnP/74Ixs3bmTjxo2UKVOGuXPn8t///peUlBRiYmIAWLRoEQEBAQA8+eSTt+3LLSVaIiIiYrjMB3THUnt7eyZOnMioUaNITU2lbNmyTJo06Y59uaVCS0RERAz3oNVZGzdutP25bt26REZG3vK4nPpyQ1OHIiIiIiZRoiUiIiKGM2t7h/xGiZaIiIiISZRoyT/uetT9LSyUB1ch/UR5aF2P6JPXQ5B8RoGWlX4sioiIiOE0dWilQkv+cZVfX5nXQxATnJjShsItlFY+jK6vG8r+U1fzehhigjoViuX1EB56KrRERETEcAq0rLQYXkRERMQkSrRERETEcFqjZaVCS0RERAynQstKU4ciIiIiJlGiJSIiIoZToGWlREtERETEJEq0RERExHBao2WlREtERETEJEq0RERExHAKtKxUaImIiIjhNHVopalDEREREZMo0RIRERHDKdCyUqIlIiIiYhIlWiIiImI4rdGyUqElIiIihlOdZaWpQxERERGTKNESERERw2nq0EqJloiIiIhJlGiJiIiI4RRoWanQEhEREcNp6tBKU4ciIiIiJlGiJSIiIoZToGWlREtERETEJEq0RERExHBao2WlQktEREQMpzrLSlOHIiIiIiZRoiUiIiKG09ShlRItEREREZMo0RIRERHDKdGyUqIlIiIiYhIlWiIiImI4BVpWKrRERETEcJo6tNLUoYiIiIhJlGiJiIiI4RRoWSnREhERETGJEi0RERExXGamIi1QoSUiIiIm0NShlaYORUREREyiREtEREQMp+0drJRoiYiIiJhEiZaIiIgYToGWlQotERERMZymDq00dSgiIiJiEiVaIiIiYjgFWlZKtERERERMokRLREREDKc1WlZKtERERERMokRLREREDKdEy0qFloiIiBhPdRagqUMRERER0yjREhEREcNp6tBKiZaIiIiISZRoieSSZ4lCfNSlDqWLFcQCLNx5inlb/wCga+NKhDSsSIbFwqbDcXwYeYTaFYoz/sWaANhhxydrj7Lu59i8ewFy1/q1r0u31rWww46vVh9getge3gl5mu6tanHxynUARn25lbW7T+TxSOVurFz6HRtXh4MdVKhUjT5DR7Fx1XJWhS0k9twZZi9Zj0vxEoA1lZn3+WT2RW/HyakQfYaOpsqjj+fxK8gflGhZKdG6jSlTpjBq1Cjb802bNuHl5cWxY8dsbb169WLx4sWG3XPhwoXMmzfPsOsBhISE8OSTT3L58mVbW1RUFF5eXoSGhub6uv7+/hw9etSIIeZb6ZkWxkUcpkXoVoI/2c4rDStSzaMoDaqVovmTHrSetI2WoVuZvel3AH49f5Wgj7fTZvKPdJ0VzbgXauJgb5fHr0LupEal0nRrXYvGA77Ft/c8WvlVpcoj1l/C05btoUGf+TToM19FVj6REB/H6uXfM+Gzr/lo9g9kZmayY9M6vJ6szbuhn+Pm4Znl+P3R27lw9jSfzgujx+B3mDt1Qh6NPP+xWCymPPIbFVq30aBBA6Kjo23Po6OjqV27tq0tIyODPXv24OfnZ8j90tPTeemll3j11VcNud5fPfbYY6xcudL2fNmyZTzxxBOG3+ff5mJiKofOJAKQnJrBb7FJlCleiP80rMCMDb+RlpEJwKWkNABSbmSSkWn9IeHkqG+9/OLx8q7sPnKe66npZGRa2Pbzado3fCyvhyX3ITMjg7TUVDIy0klLTaFkKTcqV3sc9zKPZDt2984tNGneGjs7Ox6rUZPkpKv8eSk+D0Yt+ZWmDm/D29ubM2fOEB8fT+nSpdm9ezf9+/dn2bJlvPzyyxw+fJiiRYtSoUIF4uLiGDt2LOfOnSM1NZU2bdrQu3dvAEJDQ4mOjubGjRuULFmS8ePHU7ZsWc6cOUPHjh0JDg5m165dvPjii8THx3Pt2jWGDRvGsmXLWLFiBS4uLhw7doxixYoxbdo03NzcSEtL44MPPiA6OhpXV1eqV69OfHw8U6dOveVrad++PeHh4bz88sskJyezZ88e2rRpQ1qatQD49ddfef/997l+/Tqpqam8+OKLtoLv+++/Z968eRQsWJDMzEw++eQTqlatmuX6X375JVu2bGH69OkUK1bMvC/KA6xsycLUKFec/Scv83bQ49Sv4sqQ1l6k3shkfMQvHDh9BYA6FUoQ+lItypYszBvf7bcVXvLgOvRHPKO7Nca1WCGup6UTUL8Ke49e4NLV6/QO8qZL8yfYe/QCw2dt4nJSal4PV+7AtbQ7gc//h74vB1LQyYla9RpQ26fBbY//M/4ipdzL2J6XKu1BQnwcJUuV/ieGm6/lx/TJDPpr9W0UKlSIWrVqER0dTVJSEtevX6dx48YcOXIEsCZcvr6+AAwbNoyQkBCWLFnC0qVL2bp1K9u3bwegR48eLF26lIiICAIDA5k8ebLtHpcvX6ZmzZqEhYXx0ksvZRvDzz//zLBhw1i5ciXVqlXj22+/BazFz7lz51i5ciXz5s3j4MGDOb6W8uXL4+TkxPHjx1mzZg3NmzfH0fF/NXbZsmWZN28eYWFhLF68mB9++IHjx48DMHHiRObPn094eDhLly7lkUf+9ze+zMxMxo4dy6FDh5g9e/a/tsgqUtCBL7rV44OwwySlpuNgb0+JIgXp8MkOJkT+wvSudW3H7j91mZahW2n38Xb6NqtGQSVbD7xfTyfw0Q/RRH74AhHjn+en43FkZGYyO3I/NV6djV+feVxISOLDns/m9VDlLiRdTSRm5xamfxPBjEVrSE25zrb1q/J6WPIQU6KVA19fX6KionB2dqZevXo4ODhQsWJFjh07RnR0NC1atODatWtER0eTkJBgOy85OZnjx4/TsGFDtm7dyoIFC7h27Rrp6elZru/k5ESrVq1ue/+6devi6WldL1C7dm127NgBWNdYtWvXDkdHRxwdHWnTpg179uzJ8bW0b9+esLAwfvrpJ959913WrVtn60tJSWH06NH8+uuv2NnZERcXx5EjR6hatSoNGjRg+PDhPPvsszzzzDOUL1/edt6IESOoW7cukydPxs7u37nWyNHeji+61SN8z1nW/nwBgAuXr7PmgPXPP526QqbFgqtzQRKS02znHY9LIjktHS/PYvz8/9MueXDNX/Mz89f8DMD73RpzNv4qcZev2fq/XH2AZR8E59Xw5B78vDca9zKP4FKiJAC+jZ7l18MHaNy89S2PL1najUtxF2zPL8XH4lra/R8Za76nQAtQopUjPz8/oqOj2b17N/Xr1wegfv367Ny5kz179uDr60tmZiZ2dnYsWbKE8PBwwsPDWb9+Pa+88gpnz55lwoQJfPTRR6xYsYLx48fbpusAChcunGOB4uTkZPuzg4MDGRkZuX4tAQEBrFy5kmvXruHl5ZWl7+OPP8bNzY2wsDAiIiKoVasWqanWKZDp06czePBgrl+/ziuvvMKWLVts59WvX589e/ZkKTL/bUI71+K32CTmbvnfQuh1B2N5qlopACq7OVPAwZ6E5DTKuRa2LX4vW7IwVd2Lcibh2i2vKw8WtxJFACjvVox2jR7l+42/UMbV2dbfruGjHP5D63byg9LuZTj2y0FSU1KwWCwc3LebshUq3fZ4n6easnX9KiwWC0cP/0wR56KaNrxLWgxvpUQrB97e3pw9e5Z169YREhICgI+PD2+//TYuLi62dKdevXrMmjWLfv36AXD+/HkcHR1JSkqiQIECuLm5kZmZyaJFiwwZl6+vL5GRkbRu3ZqMjAxWr16Nu3vOf8NydnZm6NChlC6d/QfE1atX8fLywtHRkaNHjxITE0NgYCDp6emcO3eOWrVqUatWLU6dOsUvv/xC06ZNAejYsSNPPPEEr776KnPmzMHDw8OQ15df+FQuSXD9chw5l8jKIY0AmLTyVxZHnWZi59qseasJNzIyGbLgJwDqV3Gld7OqpGdkkmmB95Yc5M/kG3n5EuQuLXyvHa4uhbiRnsngaeu5kpzKx/2aUauqOxYLnIy9woBP1935QpLnHq3+JH6NmzG878vYOzhQuaoXzVsHszpsERE/fM3lhEu81bMzdXwb0vvN9/D2bci+qO0M6tqegk6F6DNk1J1vIvIXKrRy4OTkRO3atYmNjbUVETVr1iQ2NpaAgADbcZMnT2bChAm0bdsWsBY148aNw8vLi4CAAFq3bk3JkiVp2rQpMTEx9z2uzp07c+TIEdq0aUPJkiWpUqXKXZ3XuvWto/E+ffrw1ltvsWTJEipXrmxL7zIzMxk+fDhXr17Fzs4OT09P3nzzzSznBgUF4eTkxKuvvsrs2bMpV67c/b24fCTmxJ9Ufn3lLfte/25/trawmLOExZw1e1higuZvLszW9tpErevJr17s2osXu/bK0taqQ2dadeic7Vg7OzteGzjsnxraQyU/pk9msLPonciXkpKSKFq0KGlpafTp04eAgABeeOGFvB7WXbldcSL524kpbSjcYlJeD0NMcH3dUPafuprXwxAT1Klg3oeYyvYJM+W6Z7/oYMp1zaJEK5/q1q0baWlppKam8vTTT9OhQ/76H09ERB5uynGsVGjlU0buSC8iIiLmUKElIiIixvt/7d17UFTn/cfx9yKICihaKTVKIvFCJ40CRkRjvBQzBZVFLrZgGuql0Wq91F811RpD1GgMSs002g5qNMRUvIGIW6umTVNjNIFqmyimGmNTRaWosYqCgLj8/tjh1A2KSM5po/28ZpjJnmef55w9nJP98nmeXRVoASq0RERExAKaOnTR92iJiIiIWESJloiIiJhOiZaLEi0RERERiyjREhEREdMp0XJRoSUiIiKmU6HloqlDERERue/861//Yvz48URHR2O325kyZQoXL14E4MMPPyQuLo7o6GjG80L7zAAAE8NJREFUjRvH559/bvRrqK0pVGiJiIiI+Wot+mkkm83GM888w+7du3E4HAQFBZGRkYHT6eTZZ58lLS2N3bt307t3bzIyMgAabGsqFVoiIiJy3/H39ycyMtJ4HBYWxtmzZykqKsLb25vevXsDkJKSwq5duwAabGsqrdESERER01m1RqusrIyysrJ621u3bk3r1q1v2cfpdLJhwwaioqIoKSnhgQceMNratWuH0+nk0qVLDbb5+/s36XhVaImIiIjprCq03njjDVasWFFv+5QpU5g6deot+7z44ou0atWKp59+mt///veWHNftqNASERGRe8bo0aNJSEiot/12aVZ6ejonT54kMzMTDw8POnTowNmzZ432ixcv4uHhgb+/f4NtTaVCS0RERExnVaLV0BThFy1btoyioiJWrVpF8+bNAXj00UeprKzkwIED9O7dm40bNxITE3PHtqZSoSUiIiL3nePHj7Ny5Uo6d+5MSkoKAJ06deJXv/oVS5Ys4YUXXqCqqoqOHTuydOlSADw8PG7b1lQqtERERMR8/+XvK+3WrRvHjh27ZVuvXr1wOBx33dYU+noHEREREYso0RIRERHT6Z/gcVGhJSIiIqZToeWiqUMRERERiyjREhEREdMp0XJRoiUiIiJiESVaIiIiYjolWi4qtERERMR8qrMATR2KiIiIWEaJloiIiJhOU4cuSrRERERELKJES0REREynRMtFhZaIiIiYToWWi6YORURERCyiREtERERMp0TLRYmWiIiIiEWUaImIiIj5FGgBSrRERERELKNES0REREynNVouKrRERETEdCq0XDR1KCIiImIRJVoiIiJiPiVagBItEREREcso0RIRERHz1Tr/20fwlaBCS0RERMynqUNAU4ciIiIillGiJSIiIubT1CGgREtERETEMkq0RERExHxaowWo0BIREREraOoQ0NShiIiIiGWUaImIiIj5lGgBSrRERERELKNES0RERMynxfCAEi0RERERy9hqa1VyioiIiLlaDn3FknGv7fw/S8a1iqYORURExHzKcQBNHYqIiIhYRomWiIiImE9f7wAo0RIRERGxjBItERERMZ/WaAEqtERERMQKmjoENHUoIiIiYhklWiIiImI+TR0CSrRERERELKNES0RERMynNVqAEi25j7zyyiu88MILxuN33nmHkJAQjh8/bmz70Y9+xJYtW3j77bdJT08H4PTp02zatMltrKioKD755BPTju1W+7hZQUEBISEhxjHVSU1NJSQkhPLy8ibtd+vWrUybNq1JfW9m9vmwWllZGatXr3bblpqayjvvvNPoMU6ePEl4eDg1NTUA1NbW0q9fP7ff0a5du3jqqafMOWigtLSU1NRU08YD1zUQEhLC+vXrjW21tbUMGTKEyMjIJo+7fPnyetfr3bib+9UsGzZsICsry7TxwHVdPfroo1y6dMnYdrv7+W7ca/fcLdXWWvNzj1GhJfeNvn37UlhYaDwuLCwkNDTU2Hbjxg0OHjxIZGQkQ4YMYdasWQCcOXOmwSLIDI3ZR3BwMG+//TY3btwAoLi4mIqKCkuP66umrqip43Q6aco/x1pWVsZrr732pY7loYceonXr1hw5cgSATz/9lI4dO3LgwAHjOYWFhfTp0+dL7adOTU0NgYGBvPnmm6aMd7NHHnmEbdu2GY8LCgpo06aN6fu5G3dzv5qhpqaGUaNGMWbMGFPGu1n37t3ZsWOH8Xjr1q1861vfMn0/cm9SoSX3jfDwcE6fPs2FCxcA+POf/8yPf/xjCgoKAPj444/x9fXlwQcfdEt6FixYwIkTJxgxYoRb+rNz506Sk5OJioriN7/5jbH90KFDJCcnY7fbSU5O5tChQ4DrzSsxMdF43s2Pb7ePm7Vq1Yrw8HDee+89APLy8oiPj3d7Tnp6OklJScTFxTF69GjOnDkDwOeff86YMWOw2+3Y7XZeeumleuOXlJSQmJjI7373u7s4q/WlpqaSnp7OqFGjGDJkCBkZGUZbaWkpU6dONY5j5cqVAFy4cIHJkycb229+04+KiiIjI4ORI0eSlpbG8uXLmTZtGuPGjWPYsGGUlZWxZ88eUlJSSExMJDk5mQ8//NDon5OTQ1xcHHFxcSQlJXHhwgUWLFjAlStXGDFiBCkpKW7HX1payhNPPEFVVZWxbeLEiTgcjnqvtU+fPsb1U1hYyHe+8x2uX7/O1atXjW11hcCqVasYOXIkCQkJTJw4kfPnzwPw/vvvk5ycTHx8PHa73e0NOTU1lUWLFvG9732PSZMmcfr0abfCIiQkhMzMTJKSkhgyZAi7d+822nbv3k1MTAzx8fFkZmY2mHwGBQXRokULPv30U8B1bSUkJLg9Z8aMGSQmJmK325k8eTKXL18G4O9//zvJycnExcURGxvLmjVr6o1/7Ngx7Ha7W+F0J3dzv547d45p06YxcuRI7HY7mZmZxji3uyfqzmV6ejoJCQls2bLFLYXbunUr48aNY/r06QwfPpyUlBTjd1ZdXc3zzz9PdHQ0o0aNYsGCBQ0mw/Hx8eTn5wNQXl7OwYMHGTBggNv5eeqpp0hISGDYsGFuqdqmTZsYOnQoI0aMwG63c+LEiXrjr127ltGjR3PlypVGn9+vhFqnNT/3GBVact9o0aIFPXv2pLCwkKtXr3Lt2jUGDBjA0aNHgdunD2lpaXTp0oX8/HxeffVVY3tlZSWbNm1i3bp1/OIXv6C8vJzq6mqmTZvG9OnTcTgc/OQnP2HatGlUV1c3eGy328cXJSQkkJeXR21tLTt27CA2Ntatffz48eTm5rJ9+3ZiY2ONIsfhcPDggw/icDhwOBxMnjzZrd/Ro0eZMGECc+bMYdiwYQ2fyEYoKSlh/fr1bNu2jS1btvCPf/wDgJkzZxIaGmocx3e/+10AFi5cSLdu3XA4HKxZs4aMjAy3aZGrV6+Sk5NjFIiHDh0iIyODXbt2cfnyZX7961/z2muvsXXrVhYuXMj06dMBVzG7cuVK1qxZw/bt21m3bh1+fn6kpaXh5+dHfn4+GzdudDv2wMBAIiIijILz9OnTFBUVER0dXe91RkZGGsVDYWEhERERhIeHc+DAAS5evMipU6cIDw8nPz+f4uJiNm/eTF5eHgMHDuTll18GXGlSdnY227Zt4/XXXyc9Pd0oYsCVXGZnZ9eb6qzj6+tLbm4uS5YsYeHChYCrcE1LSyMzM5Nt27bRokWLO/7O4uPjycvLMwqBgQMHurU/99xzbN26FYfDQdeuXY3jyc7OJioqiu3bt/Pb3/6WkSNHuvXbv38/M2fO5JVXXrmrdO9u7tdZs2aRmppKTk4Oubm5vPvuu+zbtw+4/T0BcOnSJXr06EFeXh6jRo2qdwyHDx9m1qxZ7Nixg65duxp/UG3atImzZ8+yY8cOsrKyKCoqavC1BAUF4e3tzYkTJ9i1axdPPvkknp7/XgLdsWNHsrKyyMvLY8uWLWzevNkoqJYsWcIbb7xBfn4+ubm5PPDAA0Y/p9PJwoULOXLkCKtXr8bPz6/R51e+OrQYXu4rdQmEj48Pjz32GM2aNeOhhx7i+PHjRiLRWHUFSadOnWjdujX//Oc/qampwcvLi379+gHw+OOP4+XlxWeffWbK8UdGRjJ//nz+8Ic/0L17d9q2bevW/u6775KdnU1FRYXbNFtoaChZWVmkp6fTp08fnnjiCaPt2LFjTJkyhZUrV9KlSxdTjjMmJgYPDw/8/Pzo0qULp06dIiAggL/+9a+8/vrrxvPatWsHuFKd2bNnA/D1r3+dQYMGUVBQQPfu3QHqJXcDBw40+u7du5dTp07x/e9/32ivqanhwoUL/OlPf2LEiBEEBAQA4OPj06jjT01NZfHixSQkJLBx40aSkpJo3rx5vedFRkby0ksvUVNTw5EjR+jRowclJSUUFBRQWVlJz549adGiBX/84x8pKioyUqIbN27g6+sLwMWLF5kzZw4nT56kWbNmXL58mc8++4ywsDAA7Ha725vyF9Vdh2FhYZw7d46qqio++ugjHnnkETp37gxAUlISixcvbvA1x8TEkJiYSOfOnXnyySdp1qyZW3t+fj4Oh4Pr169TUVFhjB0REcHSpUu5du0akZGR9O3b1+jz3nvvsXfvXtasWUNgYGCD+7+VxtyvFRUVFBYWcvHiRaNfeXk5J06coH///re9JwC8vb0ZOnTobfffq1cvOnToALjuof379wOuAn7EiBF4enri6enJ8OHDOXjwYIOvpa6Q/eijj5g7dy5vvfWW0VZZWcm8efM4duwYNpuNc+fOcfToUbp06ULfvn2ZPXs23/72txk8eDBBQUFGvzlz5tCrVy8yMjKw2WyNP7FfFc57bz2VFVRoyX2lrlDx8/MjIiICcL1RvP/++xw8eJC5c+c2eixvb2/jv5s1a2asnbqdZs2aua0nunlqqrFsNhtDhw5l7ty59d44z5w5w+LFi8nJySEoKIi//OUvzJw5E3BNw+Tl5bF//37y8/NZtWoVGzZsAFwJTnl5OQUFBaYVWnd7bu6kVatWbo+/WDANGDCAJUuWfKl93KxXr17GGqC8vDxycnJu+bygoCDatGljJIaenp5ERESwdu1aqqqqjMSltraWSZMm1Ut7AObNm0dUVBQrVqzAZrMRHR3tdm188bV/Ud25riuMvlhMNJaPjw9hYWFkZGSwbt06t7YDBw6wYcMGNm7cSLt27XA4HGzevBmA6OhowsLC2LdvH6tXryY3N9dIjYKDgzl+/DhFRUVNKrQac786nU5sNhs5OTl4eXm59W/ongBo2bJlgwWKmddxTEwMsbGxtGvXjpCQELdCa9myZQQEBPDyyy/j6enJuHHjjGtgxYoVHD58mA8++IAf/OAHzJs3j0GDBhnnoq7I/NrXvtbkY5P/Lk0dyn0lPDycM2fO8NZbbxlvgr1792b9+vW0bt3a7a/FOr6+vsaamzsJDg7m+vXrfPDBB4ArqampqSE4OJigoCCKi4u5fPmyMfXXlH0kJyfzzDPP1JvauXr1Kl5eXgQEBOB0Ot2mxIqLi/H19WX48OH8/Oc/58iRIzidrrUM/v7+ZGVlsX37dre0yWw+Pj6Eh4e7rT+pSyH69etnvHGfP3+ePXv2uCUjDenfvz979+51+zRa3bq4wYMHk5+fb6zzKS8vp6qqCl9fXyorKxssSlJTU/npT39KeHi4kWrcSp8+fcjMzDSup4CAACoqKtizZ4+xnioqKors7GxjSrC6utqYArty5QodO3bEZrOxb98+Tp482ajX3ZDQ0FA+/vhjTp06BbjWXDXG+PHjmTp1KiEhIW7by8rK8PX1xd/fn+rqanJzc422kydPEhAQQGJiIpMnT+bw4cNGW8eOHVm7di3Lli1r0tq/xtyvvr6+PPbYY6xatcroV1JSwvnz5xu8J76MPn364HA4qKmpoaqqip07d96xj4+PD88++6zxIZubXblyhW984xt4enryySefGB+oqKmpobi4mJ49ezJhwgT69+/P3/72N6NfUlISY8eOZcyYMZSWlpry2v6jtEYLUKIl9xlvb29CQ0MpLS01/sLu0aMHpaWlxMTE3LJPSEgIwcHBxMbG8vDDDze4hqp58+a8+uqrLFq0iIqKClq1asUvf/lLmjdvTmBgIGPHjiUxMZH27dsTERFhFAd3s4/AwEDGjx9/y+OMiYlh2LBhtG3blkGDBhn/wy4sLCQrKwsPDw+cTifz58/Hw+Pff0f5+fmxZs0aJk6cSEVFRb01XGbJyMhg/vz5xMbG4uHhQWxsLBMmTGDu3LmkpaVht9sB11qubt26NWrMzp07s3TpUp577jkqKyu5fv06vXr1omfPnkRGRjJhwgTGjh2LzWajefPmZGZm0r59e2PhfZs2bW75Bjx8+HAWLFhwx69niIyMZNu2bUbiAq5ELD8/35j+i4+P59KlSzz99NOAK+EaNWoU3/zmN5kxYwbz589n+fLl9OjRo16R0xTt27dn3rx5jB8/npYtWzJ48GC8vLxo2bJlg/26du1K165d620fMGAA27dvJzo6mrZt29K7d2+joNq5cycOhwMvLy9sNhtz5sxx69uhQweysrL44Q9/SGVlpdsHQu6ksfdrRkYGixcvNq4fHx8fFi1a1OA98WWkpKRw9OhRhg8fTtu2bXn44Ycb1e926x8nTZrEz372M3JycggODjauJafTyezZs7ly5Qo2m40OHTowY8YMt75xcXF4e3szZswYVq9eTadOnb7ci/tPugeLIivYapvy2WkRkXvcgQMHmDdvHg6H455c/3L16lVjHVhubi45OTnGdLF8eXXnt7q6mkmTJhETE2N8uEMap+WANEvGvbZ3gSXjWkWJloj8z5kzZw779+8nPT39niyyAN5880127drFjRs3aNOmjfGJRDHH2LFjqa6upqqqiscff7ze12FIIyjHAZRoiYiIiAVaPvG8JeNee+9FS8a1ihItERERMZ/WaAEqtERERMQKmjAD9PUOIiIiIpZRoiUiIiLm09QhoERLRERExDJKtERERMR8WqMFqNASERERK2jqENDUoYiIiIhllGiJiIiI+TR1CCjREhEREbGMEi0RERExn9ZoAUq0RERERCyjREtERETMpzVagAotERERsYKmDgFNHYqIiIhYRomWiIiImE9Th4ASLRERERHLKNESERER82mNFgC22lpleyIiIiJW0NShiIiIiEVUaImIiIhYRIWWiIiIiEVUaImIiIhYRIWWiIiIiEVUaImIiIhY5P8BbnHpLU+KLocAAAAASUVORK5CYII=\n",
            "text/plain": [
              "<Figure size 576x576 with 2 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "5dZs2eMfODa8"
      },
      "source": [
        "#### MultinomialNB w/o Hyperparameter Tuning"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "lILXKzSUXZDi"
      },
      "source": [
        "# MultinomialNB Building\n",
        "multinomialNB_model = sklearn.naive_bayes.MultinomialNB()"
      ],
      "execution_count": 13,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "vpDHxgL_agbw",
        "outputId": "cf5ca9ca-c465-4e77-a679-e03916230f29"
      },
      "source": [
        "# Training\n",
        "multinomialNB_model.fit(X_train, y_train)"
      ],
      "execution_count": 14,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "MultinomialNB(alpha=1.0, class_prior=None, fit_prior=True)"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 14
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 602
        },
        "id": "MJOD5NBjbc4_",
        "outputId": "fdbe0ad6-91ca-4396-9a2c-2cbbab5d1190"
      },
      "source": [
        "# Evaluate\n",
        "mnb_y_pred = multinomialNB_model.predict(X_test)\n",
        "\n",
        "print('MultinomialNB Metrics', end='\\n\\n')\n",
        "accuracy = sklearn.metrics.accuracy_score(y_test, mnb_y_pred)\n",
        "print('Accuracy: %f' % accuracy, end='\\n\\n')\n",
        "\n",
        "recall = sklearn.metrics.recall_score(y_test, mnb_y_pred, average='micro')\n",
        "print('Recall: %f' % recall, end='\\n\\n')\n",
        "\n",
        "f1_score = sklearn.metrics.f1_score(y_test, mnb_y_pred, average='micro')\n",
        "print('F1 Score: %f' % f1_score)\n",
        "\n",
        "cm = sklearn.metrics.confusion_matrix(y_test, mnb_y_pred)\n",
        "# print('Confusion Matrix:')\n",
        "# print(cm, end='\\n\\n')\n",
        "plt.figure(figsize=(8,8), facecolor='w')\n",
        "sns.heatmap(cm, annot=True, fmt='d', linewidths=.5, square=True,cmap='Blues_r', yticklabels=CATEGORIES, xticklabels=CATEGORIES)\n",
        "plt.title('Confusion Matrix for MultinomialNB')\n",
        "\n",
        "print()"
      ],
      "execution_count": 15,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "MultinomialNB Metrics\n",
            "\n",
            "Accuracy: 0.687394\n",
            "\n",
            "Recall: 0.687394\n",
            "\n",
            "F1 Score: 0.687394\n",
            "\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 576x576 with 2 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "sgBAuy-hOJMc"
      },
      "source": [
        "#### GaussianNB with Hyperparameter tuning"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "SOkbbPqUVC6-",
        "outputId": "812f5a82-c547-4f1b-bca4-3a333767679e"
      },
      "source": [
        "%%time\n",
        "\n",
        "# Gridsearch for GaussianNB\n",
        "params = {\n",
        "    'var_smoothing' : [1e-09, 1e-07, 1e-05, 1e-03, 1, 100]\n",
        "}\n",
        "\n",
        "# Hyperparameter tuning\n",
        "grid_search_cv = sklearn.model_selection.GridSearchCV(sklearn.naive_bayes.GaussianNB(), param_grid=params, scoring='accuracy', n_jobs=-1)\n",
        "\n",
        "grid_search_cv.fit(X_train, y_train)"
      ],
      "execution_count": 16,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "CPU times: user 870 ms, sys: 407 ms, total: 1.28 s\n",
            "Wall time: 23.4 s\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "4reLm42VYTs-",
        "outputId": "00cc756e-e571-42f3-9e74-0c74c9222db3"
      },
      "source": [
        "print('Best param for \\'var_smoothing\\':', grid_search_cv.best_params_)\n",
        "print('With Accuracy of %f' % grid_search_cv.best_score_)"
      ],
      "execution_count": 17,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Best param for 'var_smoothing': {'var_smoothing': 0.001}\n",
            "With Accuracy of 0.771927\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 602
        },
        "id": "YkEqKFirYg7H",
        "outputId": "d3c8bb48-a307-4655-f6d5-fac034ab0df4"
      },
      "source": [
        "# Grab the best performing model\n",
        "gnb_model = grid_search_cv.best_estimator_\n",
        "\n",
        "gnb_gs_y_pred = gnb_model.predict(X_test)\n",
        "\n",
        "print('GaussianNB with \\'var_smoothing\\' of %f Metrics' % grid_search_cv.best_params_['var_smoothing'], end='\\n\\n')\n",
        "accuracy = sklearn.metrics.accuracy_score(y_test, gnb_gs_y_pred)\n",
        "print('Accuracy: %f' % accuracy, end='\\n\\n')\n",
        "\n",
        "recall = sklearn.metrics.recall_score(y_test, gnb_gs_y_pred, average='micro')\n",
        "print('Recall: %f' % recall, end='\\n\\n')\n",
        "\n",
        "f1_score = sklearn.metrics.f1_score(y_test, gnb_gs_y_pred, average='micro')\n",
        "print('F1 Score: %f' % f1_score)\n",
        "\n",
        "cm = sklearn.metrics.confusion_matrix(y_test, gnb_gs_y_pred)\n",
        "# print('Confusion Matrix:')\n",
        "# print(cm, end='\\n\\n')\n",
        "plt.figure(figsize=(8,8), facecolor='w')\n",
        "sns.heatmap(cm, annot=True, fmt='d', linewidths=.5, square=True,cmap='Blues_r', yticklabels=CATEGORIES, xticklabels=CATEGORIES)\n",
        "plt.title('Confusion Matrix for GaussianNB')\n",
        "\n",
        "print()"
      ],
      "execution_count": 18,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "GaussianNB with 'var_smoothing' of 0.001000 Metrics\n",
            "\n",
            "Accuracy: 0.778689\n",
            "\n",
            "Recall: 0.778689\n",
            "\n",
            "F1 Score: 0.778689\n",
            "\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 576x576 with 2 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "3vPRDlBiOPC9"
      },
      "source": [
        "#### MultinomialNB with Hyperparameter tuning"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "Nj9s45_calLq",
        "outputId": "ac4c5cd5-3243-4faa-d74b-0f1287662170"
      },
      "source": [
        "# Hyperparameter tuning for multinomialnb\n",
        "params = {\n",
        "    'alpha' : [0.1, 0.5, 1, 2, 5, 100],\n",
        "}\n",
        "grid_search_cv = sklearn.model_selection.GridSearchCV(\n",
        "    sklearn.naive_bayes.MultinomialNB(),\n",
        "    param_grid=params,\n",
        "    scoring='accuracy',\n",
        "    n_jobs=-1\n",
        ")\n",
        "\n",
        "grid_search_cv.fit(X_train, y_train)\n"
      ],
      "execution_count": 19,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "GridSearchCV(cv=None, error_score=nan,\n",
              "             estimator=MultinomialNB(alpha=1.0, class_prior=None,\n",
              "                                     fit_prior=True),\n",
              "             iid='deprecated', n_jobs=-1,\n",
              "             param_grid={'alpha': [0.1, 0.5, 1, 2, 5, 100]},\n",
              "             pre_dispatch='2*n_jobs', refit=True, return_train_score=False,\n",
              "             scoring='accuracy', verbose=0)"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 19
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "0-BlDp5Gktcb",
        "outputId": "bec8727b-c8b7-4075-c80c-0845ecd33f5d"
      },
      "source": [
        "print('Best params:', grid_search_cv.best_params_)\n",
        "print('With Accuracy of %f ' % grid_search_cv.best_score_)"
      ],
      "execution_count": 20,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Best params: {'alpha': 100}\n",
            "With Accuracy of 0.681816 \n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 602
        },
        "id": "lXSQ7EIXl4Tt",
        "outputId": "9fa121d2-09a5-4969-f72c-1c13f1deef1b"
      },
      "source": [
        "mnb_model = grid_search_cv.best_estimator_\n",
        "\n",
        "mnb_gs_y_pred = mnb_model.predict(X_test)\n",
        "\n",
        "print('MultinomialNB with \\'alpha\\' of %f Metrics' % grid_search_cv.best_params_['alpha'], end='\\n\\n')\n",
        "accuracy = sklearn.metrics.accuracy_score(y_test, mnb_gs_y_pred)\n",
        "print('Accuracy: %f' % accuracy, end='\\n\\n')\n",
        "\n",
        "recall = sklearn.metrics.recall_score(y_test, mnb_gs_y_pred, average='micro')\n",
        "print('Recall: %f' % recall, end='\\n\\n')\n",
        "\n",
        "f1_score = sklearn.metrics.f1_score(y_test, mnb_gs_y_pred, average='micro')\n",
        "print('F1 Score: %f' % f1_score)\n",
        "\n",
        "cm = sklearn.metrics.confusion_matrix(y_test, mnb_gs_y_pred)\n",
        "# print('Confusion Matrix:')\n",
        "# print(cm, end='\\n\\n')\n",
        "plt.figure(figsize=(8,8), facecolor='w')\n",
        "sns.heatmap(cm, annot=True, fmt='d', linewidths=.5, square=True,cmap='Blues_r', yticklabels=CATEGORIES, xticklabels=CATEGORIES)\n",
        "plt.title('Confusion Matrix for GaussianNB')\n",
        "\n",
        "print()"
      ],
      "execution_count": 21,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "MultinomialNB with 'alpha' of 100.000000 Metrics\n",
            "\n",
            "Accuracy: 0.687394\n",
            "\n",
            "Recall: 0.687394\n",
            "\n",
            "F1 Score: 0.687394\n",
            "\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 576x576 with 2 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "PBdAewobOUeu"
      },
      "source": [
        "#### Conclusion"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "x6RqWX-tO0Ma"
      },
      "source": [
        "| Metrics | GaussianNB | MultinomialNB | GaussianNB w/'var_smoothing' of 0.001 | MultinomialNB w/'alpha' of 100\n",
        "   | --- | --- | --- | --- | --- |\n",
        "   | Accuracy | 0.779536 | 0.687394 | 0.778689 | 0.687394 |\n",
        "   | Recall   | 0.779536 | 0.687394 | 0.778689 | 0.687394 |\n",
        "   | F1 Score | 0.779536 | 0.687394 | 0.778689 | 0.687394 |\n",
        "\n",
        "   -----------------------------------------\n"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "yk-d2MGKPDCQ"
      },
      "source": [
        ""
      ],
      "execution_count": 21,
      "outputs": []
    }
  ]
}