Capstone_final2.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "Capstone_final2.ipynb",
      "provenance": [],
      "collapsed_sections": [],
      "toc_visible": true
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "language_info": {
      "name": "python"
    }
  },
  "cells": [
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "1-P_KRc1ktOO"
      },
      "outputs": [],
      "source": [
        "# Letter Classification \n",
        "# \n",
        "#Be sure to download the OCR_file for data\n",
        "#https://drive.google.com/file/d/1YV7VDUC2ot_nzc6vYql4fz_cRoBe1QG2/view?usp=sharing\n",
        "#\n",
        "#"
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "#Link for Dataset\n",
        "[OCR DATA](https://drive.google.com/file/d/1YV7VDUC2ot_nzc6vYql4fz_cRoBe1QG2/view?usp=sharing)"
      ],
      "metadata": {
        "id": "q7YCjPFG7rwy"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "# Binary Image Classification\n",
        "Data! Great, well, what kind of data?\n"
      ],
      "metadata": {
        "id": "6aF1_5C2rNwX"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "## Import libraries, libraries galore"
      ],
      "metadata": {
        "id": "SyDxNkz97L_B"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "import pandas as pd\n",
        "import numpy\n",
        "import numpy as np\n",
        "import matplotlib.pyplot as plt\n",
        "from sklearn.naive_bayes import MultinomialNB\n",
        "from sklearn.metrics import confusion_matrix\n",
        "import itertools\n",
        "from sklearn.metrics import precision_recall_fscore_support\n",
        "from sklearn.metrics import classification_report\n",
        "from sklearn.metrics import roc_curve, auc\n",
        "from sklearn.metrics import roc_auc_score\n",
        "from sklearn.svm import SVC\n",
        "from sklearn.model_selection import GridSearchCV\n",
        "from sklearn.naive_bayes import ComplementNB\n",
        "from sklearn.naive_bayes import CategoricalNB\n",
        "\n"
      ],
      "metadata": {
        "id": "AzSpkUd5reUv"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "# Data Visualization"
      ],
      "metadata": {
        "id": "sicFvyTQGNZg"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "! unzip ocr_data.zip\n",
        "! ls\n",
        "\n",
        "ocr_test = pd.read_table('ocr10-test.txt', sep=',')\n",
        "ocr_train = pd.read_table('ocr10-train.txt', sep=',')\n",
        "\n",
        "#Brought the two .txt files into the playing field\n",
        "#No benchwarmers over here \n",
        "#5 rows x 129 col\n",
        "#Features are numbered from int '0' to remaining floats '0.111'\n",
        "#ie 0, 0.1, 0.2,...\n",
        "\n",
        "\n",
        "train_features = ocr_train.columns[:-1]\n",
        "test_features = ocr_test.columns[:-1]\n",
        "#because the last row is the target column\n",
        "#make it so\n",
        "\n",
        "train_target = ocr_train.columns[-1]\n",
        "test_target = ocr_test.columns[-1]\n",
        "\n",
        "#convert to numpy for visuals\n",
        "ocr_train_np=ocr_train[train_features].to_numpy()\n",
        "ocr_test_np=ocr_test[test_features].to_numpy()\n",
        "\n",
        "#toprint\n",
        " \n",
        "# reshape because the original images are 16x8\n",
        "letter = ocr_train_np[0,:].reshape(16,8)\n",
        "plt.imshow(letter)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 354
        },
        "id": "tWJ7hnWpPh1S",
        "outputId": "d8e9e673-393e-4200-c4a6-a137323d0eb5"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Archive:  ocr_data.zip\n",
            "  inflating: ocr10-test.txt          \n",
            "  inflating: ocr10-train.txt         \n",
            "ocr10-test.txt\tocr10-train.txt  ocr_data.zip  sample_data\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<matplotlib.image.AxesImage at 0x7fac64b9a510>"
            ]
          },
          "metadata": {},
          "execution_count": 4
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "# MultiNominal Naive Bayes Model"
      ],
      "metadata": {
        "id": "XDJiITomQext"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "#lets train naive bayes model\n",
        "#MulitNominal Model\n",
        "from sklearn.naive_bayes import MultinomialNB\n",
        "\n",
        "# assign the training and test data to variables\n",
        "features = ocr_train[train_features]\n",
        "target = ocr_train[train_target]\n",
        "\n",
        "features_test = ocr_test[test_features]\n",
        "target_test=ocr_test[test_target]\n",
        "\n",
        "# fit the multinomial Naive Bayes model\n",
        "mnm = MultinomialNB(alpha=1.0, fit_prior=True)\n",
        "mnm.fit(features, target)\n",
        "\n",
        "# predict outcome for the test data\n",
        "prediction = mnm.predict(features_test)\n"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "Q72_I91TQo_G",
        "outputId": "14861496-df40-4738-c87c-24f254b1ceb8"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "/usr/local/lib/python3.7/dist-packages/sklearn/base.py:493: FutureWarning: The feature names should match those that were passed during fit. Starting version 1.2, an error will be raised.\n",
            "Feature names unseen at fit time:\n",
            "- 1.17\n",
            "- 1.18\n",
            "- 1.19\n",
            "Feature names seen at fit time, yet now missing:\n",
            "- 0.108\n",
            "- 0.109\n",
            "- 0.110\n",
            "\n",
            "  warnings.warn(message, FutureWarning)\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "# Confusion Matrix"
      ],
      "metadata": {
        "id": "-iuwot4uRIpC"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "Confusion matrix is a matrix that allows you to visualize the performance of the classification machine learning models. With this visualization, you can get a better idea of how your machine learning model is performing."
      ],
      "metadata": {
        "id": "i96KKn0qxhMz"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "\n",
        "t_features=ocr_test[test_features]\n",
        "t_target = ocr_test[test_target]\n",
        "t_target = t_target.to_numpy()\n",
        "\n",
        "\n",
        "cMatrix = confusion_matrix(t_target, prediction, labels=range(10))\n",
        "print(cMatrix)\n"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "W5BUieiSRc7u",
        "outputId": "cb466370-9093-4ec1-b75b-54f32894f8a5"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[[1163    2   29   33   61    3   33   17    1    2]\n",
            " [  10  295    0   14   12    3    6   59   28    1]\n",
            " [ 129    7  450    6   91    5   15    1    1    0]\n",
            " [  54    6    4  334   16    2    7    2   40   16]\n",
            " [ 378   17  109    6 1025   69   46    2    0    0]\n",
            " [   9    4    8   13   29  202    2    8   32    0]\n",
            " [  70   17    4   76   40   18  567    6   26    0]\n",
            " [  11   46    1   12    4    0    3  185   25    0]\n",
            " [  15   21   12   90    6   70   20   35 1311   58]\n",
            " [   1    2    0    6    1    0    7    0   16   30]]\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "# Display Confusion Matrix as an image "
      ],
      "metadata": {
        "id": "WYEKZvV2RoKV"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "matplotlib.pyplot.imshow, the input may either be actual RGB(A) data, or 2D scalar data, which will be rendered as a pseudocolor image."
      ],
      "metadata": {
        "id": "pmO_DjR3SFbg"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "classes = ['a','b','c','d','e','f','g','h','i','j']\n",
        "\n",
        "plt.figure(figsize=(10,10))\n",
        "\n",
        "plt.imshow(cMatrix, interpolation='nearest', cmap=plt.cm.Blues_r)\n",
        "plt.title('Confusion Matrix - MultiNominalNB Model w/ default hyperparameters')\n",
        "plt.colorbar()\n",
        "plt.xticks(range(10),classes)\n",
        "plt.yticks(range(10),classes)\n",
        "\n",
        "for k,v in itertools.product(range(cMatrix.shape[0]),range(cMatrix.shape[1])):\n",
        "  plt.text(v, k, cMatrix[k,v],horizontalalignment=\"center\",color=\"white\" if cMatrix[k,v]>800 else \"black\")\n",
        "\n",
        "plt.show()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 580
        },
        "id": "v203ATKfSGXW",
        "outputId": "deafc789-c792-424f-ea0f-15b0b2395ff8"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 720x720 with 2 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "# Precision recall fscore support "
      ],
      "metadata": {
        "id": "dGRkuYSGSm6P"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "\n",
        "y_true = np.array(t_target)\n",
        "y_pred = np.array(prediction)\n",
        "\n",
        "\n",
        "precision_recall_fscore_support(y_true, y_pred, average='macro')\n",
        "\n",
        "precision_recall_fscore_support(y_true, y_pred, average='micro')\n",
        "\n",
        "precision_recall_fscore_support(y_true, y_pred, average='weighted')"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "6xyHkiEOSt2q",
        "outputId": "0a80f37c-492f-44c1-b115-f15745bcc9ec"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(0.7403475582473816, 0.7196273774097555, 0.7223662318201954, None)"
            ]
          },
          "metadata": {},
          "execution_count": 8
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "# Classification Report"
      ],
      "metadata": {
        "id": "1W3Bpo7NS5KV"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "Precision, recall, f1-score and support all in one report"
      ],
      "metadata": {
        "id": "lJ6s2LlSS-VO"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "print(\"MultiNominalNB Classification Report w/ default hyperparameters\")\n",
        "print(classification_report(y_true, y_pred, target_names=classes))"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "6NMeru1uUURa",
        "outputId": "6d5948dc-98f5-4312-eaff-8030691d5fa7"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "MultiNominalNB Classification Report w/ default hyperparameters\n",
            "              precision    recall  f1-score   support\n",
            "\n",
            "           a       0.63      0.87      0.73      1344\n",
            "           b       0.71      0.69      0.70       428\n",
            "           c       0.73      0.64      0.68       705\n",
            "           d       0.57      0.69      0.62       481\n",
            "           e       0.80      0.62      0.70      1652\n",
            "           f       0.54      0.66      0.59       307\n",
            "           g       0.80      0.69      0.74       824\n",
            "           h       0.59      0.64      0.61       287\n",
            "           i       0.89      0.80      0.84      1638\n",
            "           j       0.28      0.48      0.35        63\n",
            "\n",
            "    accuracy                           0.72      7729\n",
            "   macro avg       0.65      0.68      0.66      7729\n",
            "weighted avg       0.74      0.72      0.72      7729\n",
            "\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "Hmm, we'll work on the hyper parameters shortly. Let's run the ROC. Receiver Operating Characteristic (ROC) metric to evaluate classifier output quality. ROC curves are typically used in binary classification to study the output of a classifier. Get it? Got it? alright, here we go. "
      ],
      "metadata": {
        "id": "nstd2AJxVYel"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "# Receiver Operating Characteristic (ROC)"
      ],
      "metadata": {
        "id": "f_Ys9TjTVg3C"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "A receiver operating characteristic curve, or ROC curve, is a graphical plot that illustrates the diagnostic ability of a binary classifier system as its discrimination threshold is varied. Let's create a ROC curve with *our* selected binary clasifier system."
      ],
      "metadata": {
        "id": "tHEONSoCYzHM"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "#The ROC curve is created by plotting the true positive rate (TPR) against the \n",
        "#false positive rate (FPR) at various threshold settings. The true-positive rate\n",
        "# is also known as sensitivity, recall or probability of detection.\n",
        "\n",
        "\n",
        "lw = 2\n",
        "plt.figure()\n",
        "\n",
        "\n",
        "#range represents classes a-j\n",
        "for rocClass in range(10):\n",
        "  predict_prob = mnm.predict_proba(features_test)\n",
        "  prob_possibility = predict_prob[:, rocClass]\n",
        "  thresholds = np.arange(0.0, 1.2, 0.1) #just as range gives a range of numbers and its skip patter so does numpy arange\n",
        "  true_pos, false_pos = [0] * len(thresholds), [0] * len(thresholds)\n",
        "  for pred, y in zip(prob_possibility, target_test):\n",
        "    for i, threshold in enumerate(thresholds):\n",
        "      if pred >= threshold:\n",
        "        if y == rocClass:\n",
        "          true_pos[i] += 1\n",
        "        else:\n",
        "          false_pos[i] += 1\n",
        "      else:\n",
        "        break\n",
        "  true_pos_rate = [tp/len(np.where(target_test == rocClass)[0]) for tp in true_pos]\n",
        "  false_pos_rate = [fp/len(np.where(target_test != rocClass)[0]) for fp in false_pos]\n",
        "  roc_score = roc_auc_score(target_test == rocClass, prob_possibility, max_fpr=1)\n",
        "  plt.plot(false_pos_rate, true_pos_rate, label='ROC curve of class {0} (area = {1:0.2f})'.format(classes[rocClass], roc_score))\n",
        "  plt.legend(loc=\"lower right\")\n",
        "  plt.xlim([0.0,1.0])\n",
        "plt.ylim([0.0, 1.0])\n",
        "plt.xlabel('False Positive Rate')\n"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        },
        "id": "NDrf01tzVjbQ",
        "outputId": "449492f4-8546-40d0-bdb1-3afafaac717e"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "/usr/local/lib/python3.7/dist-packages/sklearn/base.py:493: FutureWarning: The feature names should match those that were passed during fit. Starting version 1.2, an error will be raised.\n",
            "Feature names unseen at fit time:\n",
            "- 1.17\n",
            "- 1.18\n",
            "- 1.19\n",
            "Feature names seen at fit time, yet now missing:\n",
            "- 0.108\n",
            "- 0.109\n",
            "- 0.110\n",
            "\n",
            "  warnings.warn(message, FutureWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/base.py:493: FutureWarning: The feature names should match those that were passed during fit. Starting version 1.2, an error will be raised.\n",
            "Feature names unseen at fit time:\n",
            "- 1.17\n",
            "- 1.18\n",
            "- 1.19\n",
            "Feature names seen at fit time, yet now missing:\n",
            "- 0.108\n",
            "- 0.109\n",
            "- 0.110\n",
            "\n",
            "  warnings.warn(message, FutureWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/base.py:493: FutureWarning: The feature names should match those that were passed during fit. Starting version 1.2, an error will be raised.\n",
            "Feature names unseen at fit time:\n",
            "- 1.17\n",
            "- 1.18\n",
            "- 1.19\n",
            "Feature names seen at fit time, yet now missing:\n",
            "- 0.108\n",
            "- 0.109\n",
            "- 0.110\n",
            "\n",
            "  warnings.warn(message, FutureWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/base.py:493: FutureWarning: The feature names should match those that were passed during fit. Starting version 1.2, an error will be raised.\n",
            "Feature names unseen at fit time:\n",
            "- 1.17\n",
            "- 1.18\n",
            "- 1.19\n",
            "Feature names seen at fit time, yet now missing:\n",
            "- 0.108\n",
            "- 0.109\n",
            "- 0.110\n",
            "\n",
            "  warnings.warn(message, FutureWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/base.py:493: FutureWarning: The feature names should match those that were passed during fit. Starting version 1.2, an error will be raised.\n",
            "Feature names unseen at fit time:\n",
            "- 1.17\n",
            "- 1.18\n",
            "- 1.19\n",
            "Feature names seen at fit time, yet now missing:\n",
            "- 0.108\n",
            "- 0.109\n",
            "- 0.110\n",
            "\n",
            "  warnings.warn(message, FutureWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/base.py:493: FutureWarning: The feature names should match those that were passed during fit. Starting version 1.2, an error will be raised.\n",
            "Feature names unseen at fit time:\n",
            "- 1.17\n",
            "- 1.18\n",
            "- 1.19\n",
            "Feature names seen at fit time, yet now missing:\n",
            "- 0.108\n",
            "- 0.109\n",
            "- 0.110\n",
            "\n",
            "  warnings.warn(message, FutureWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/base.py:493: FutureWarning: The feature names should match those that were passed during fit. Starting version 1.2, an error will be raised.\n",
            "Feature names unseen at fit time:\n",
            "- 1.17\n",
            "- 1.18\n",
            "- 1.19\n",
            "Feature names seen at fit time, yet now missing:\n",
            "- 0.108\n",
            "- 0.109\n",
            "- 0.110\n",
            "\n",
            "  warnings.warn(message, FutureWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/base.py:493: FutureWarning: The feature names should match those that were passed during fit. Starting version 1.2, an error will be raised.\n",
            "Feature names unseen at fit time:\n",
            "- 1.17\n",
            "- 1.18\n",
            "- 1.19\n",
            "Feature names seen at fit time, yet now missing:\n",
            "- 0.108\n",
            "- 0.109\n",
            "- 0.110\n",
            "\n",
            "  warnings.warn(message, FutureWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/base.py:493: FutureWarning: The feature names should match those that were passed during fit. Starting version 1.2, an error will be raised.\n",
            "Feature names unseen at fit time:\n",
            "- 1.17\n",
            "- 1.18\n",
            "- 1.19\n",
            "Feature names seen at fit time, yet now missing:\n",
            "- 0.108\n",
            "- 0.109\n",
            "- 0.110\n",
            "\n",
            "  warnings.warn(message, FutureWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/base.py:493: FutureWarning: The feature names should match those that were passed during fit. Starting version 1.2, an error will be raised.\n",
            "Feature names unseen at fit time:\n",
            "- 1.17\n",
            "- 1.18\n",
            "- 1.19\n",
            "Feature names seen at fit time, yet now missing:\n",
            "- 0.108\n",
            "- 0.109\n",
            "- 0.110\n",
            "\n",
            "  warnings.warn(message, FutureWarning)\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "Text(0.5, 0, 'False Positive Rate')"
            ]
          },
          "metadata": {},
          "execution_count": 10
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "# Hyper Parameters with GridSearchCV"
      ],
      "metadata": {
        "id": "TUoi5sq3VwIa"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "#defining parameter range\n",
        "#parameter grid takes a dictionary containing list of \n",
        "#values of each hyperparameter\n",
        "\n",
        "param_grid = {'alpha': [0.1,0.2,0.3,0.4,0.5,0.6,0.7,0.8,0.9,1.0,1.1]} \n",
        "grid = GridSearchCV(mnm, param_grid, refit = True, verbose = 3)\n",
        "\n",
        "\n",
        "# fitting the model for grid search\n",
        "grid.fit(features, target)\n",
        "\n",
        "\n",
        "# print best parameter after tuning\n",
        "print(grid.best_params_)\n",
        "\n",
        "\n",
        "# print how our model looks after hyper-parameter tuning\n",
        "print(grid.best_estimator_)\n",
        "\n",
        "\n",
        "#testing the model\n",
        "grid_predictions = grid.predict(features_test)\n",
        "print(classification_report(target_test, grid_predictions))"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "2lSdnEhQV5aT",
        "outputId": "32de074a-2f94-41ce-a734-01a746a97f99"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Fitting 5 folds for each of 11 candidates, totalling 55 fits\n",
            "[CV 1/5] END .........................alpha=0.1;, score=0.706 total time=   0.1s\n",
            "[CV 2/5] END .........................alpha=0.1;, score=0.698 total time=   0.1s\n",
            "[CV 3/5] END .........................alpha=0.1;, score=0.732 total time=   0.1s\n",
            "[CV 4/5] END .........................alpha=0.1;, score=0.731 total time=   0.1s\n",
            "[CV 5/5] END .........................alpha=0.1;, score=0.729 total time=   0.1s\n",
            "[CV 1/5] END .........................alpha=0.2;, score=0.706 total time=   0.1s\n",
            "[CV 2/5] END .........................alpha=0.2;, score=0.698 total time=   0.1s\n",
            "[CV 3/5] END .........................alpha=0.2;, score=0.732 total time=   0.1s\n",
            "[CV 4/5] END .........................alpha=0.2;, score=0.731 total time=   0.1s\n",
            "[CV 5/5] END .........................alpha=0.2;, score=0.729 total time=   0.1s\n",
            "[CV 1/5] END .........................alpha=0.3;, score=0.706 total time=   0.1s\n",
            "[CV 2/5] END .........................alpha=0.3;, score=0.698 total time=   0.1s\n",
            "[CV 3/5] END .........................alpha=0.3;, score=0.732 total time=   0.1s\n",
            "[CV 4/5] END .........................alpha=0.3;, score=0.731 total time=   0.1s\n",
            "[CV 5/5] END .........................alpha=0.3;, score=0.729 total time=   0.1s\n",
            "[CV 1/5] END .........................alpha=0.4;, score=0.706 total time=   0.1s\n",
            "[CV 2/5] END .........................alpha=0.4;, score=0.698 total time=   0.1s\n",
            "[CV 3/5] END .........................alpha=0.4;, score=0.732 total time=   0.1s\n",
            "[CV 4/5] END .........................alpha=0.4;, score=0.731 total time=   0.1s\n",
            "[CV 5/5] END .........................alpha=0.4;, score=0.729 total time=   0.1s\n",
            "[CV 1/5] END .........................alpha=0.5;, score=0.706 total time=   0.1s\n",
            "[CV 2/5] END .........................alpha=0.5;, score=0.698 total time=   0.1s\n",
            "[CV 3/5] END .........................alpha=0.5;, score=0.732 total time=   0.1s\n",
            "[CV 4/5] END .........................alpha=0.5;, score=0.731 total time=   0.1s\n",
            "[CV 5/5] END .........................alpha=0.5;, score=0.729 total time=   0.1s\n",
            "[CV 1/5] END .........................alpha=0.6;, score=0.706 total time=   0.1s\n",
            "[CV 2/5] END .........................alpha=0.6;, score=0.698 total time=   0.1s\n",
            "[CV 3/5] END .........................alpha=0.6;, score=0.732 total time=   0.1s\n",
            "[CV 4/5] END .........................alpha=0.6;, score=0.731 total time=   0.1s\n",
            "[CV 5/5] END .........................alpha=0.6;, score=0.730 total time=   0.1s\n",
            "[CV 1/5] END .........................alpha=0.7;, score=0.707 total time=   0.1s\n",
            "[CV 2/5] END .........................alpha=0.7;, score=0.698 total time=   0.1s\n",
            "[CV 3/5] END .........................alpha=0.7;, score=0.732 total time=   0.1s\n",
            "[CV 4/5] END .........................alpha=0.7;, score=0.730 total time=   0.1s\n",
            "[CV 5/5] END .........................alpha=0.7;, score=0.730 total time=   0.1s\n",
            "[CV 1/5] END .........................alpha=0.8;, score=0.707 total time=   0.1s\n",
            "[CV 2/5] END .........................alpha=0.8;, score=0.698 total time=   0.1s\n",
            "[CV 3/5] END .........................alpha=0.8;, score=0.732 total time=   0.1s\n",
            "[CV 4/5] END .........................alpha=0.8;, score=0.730 total time=   0.1s\n",
            "[CV 5/5] END .........................alpha=0.8;, score=0.730 total time=   0.1s\n",
            "[CV 1/5] END .........................alpha=0.9;, score=0.706 total time=   0.1s\n",
            "[CV 2/5] END .........................alpha=0.9;, score=0.698 total time=   0.1s\n",
            "[CV 3/5] END .........................alpha=0.9;, score=0.732 total time=   0.1s\n",
            "[CV 4/5] END .........................alpha=0.9;, score=0.731 total time=   0.1s\n",
            "[CV 5/5] END .........................alpha=0.9;, score=0.729 total time=   0.1s\n",
            "[CV 1/5] END .........................alpha=1.0;, score=0.706 total time=   0.1s\n",
            "[CV 2/5] END .........................alpha=1.0;, score=0.698 total time=   0.1s\n",
            "[CV 3/5] END .........................alpha=1.0;, score=0.732 total time=   0.1s\n",
            "[CV 4/5] END .........................alpha=1.0;, score=0.731 total time=   0.1s\n",
            "[CV 5/5] END .........................alpha=1.0;, score=0.729 total time=   0.1s\n",
            "[CV 1/5] END .........................alpha=1.1;, score=0.706 total time=   0.1s\n",
            "[CV 2/5] END .........................alpha=1.1;, score=0.698 total time=   0.1s\n",
            "[CV 3/5] END .........................alpha=1.1;, score=0.732 total time=   0.1s\n",
            "[CV 4/5] END .........................alpha=1.1;, score=0.731 total time=   0.1s\n",
            "[CV 5/5] END .........................alpha=1.1;, score=0.729 total time=   0.1s\n",
            "{'alpha': 0.8}\n",
            "MultinomialNB(alpha=0.8)\n",
            "              precision    recall  f1-score   support\n",
            "\n",
            "           0       0.63      0.86      0.73      1344\n",
            "           1       0.71      0.69      0.70       428\n",
            "           2       0.73      0.64      0.68       705\n",
            "           3       0.57      0.70      0.62       481\n",
            "           4       0.80      0.62      0.70      1652\n",
            "           5       0.54      0.66      0.59       307\n",
            "           6       0.80      0.69      0.74       824\n",
            "           7       0.59      0.64      0.61       287\n",
            "           8       0.89      0.80      0.84      1638\n",
            "           9       0.28      0.48      0.36        63\n",
            "\n",
            "    accuracy                           0.72      7729\n",
            "   macro avg       0.65      0.68      0.66      7729\n",
            "weighted avg       0.74      0.72      0.72      7729\n",
            "\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "/usr/local/lib/python3.7/dist-packages/sklearn/base.py:493: FutureWarning: The feature names should match those that were passed during fit. Starting version 1.2, an error will be raised.\n",
            "Feature names unseen at fit time:\n",
            "- 1.17\n",
            "- 1.18\n",
            "- 1.19\n",
            "Feature names seen at fit time, yet now missing:\n",
            "- 0.108\n",
            "- 0.109\n",
            "- 0.110\n",
            "\n",
            "  warnings.warn(message, FutureWarning)\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "We found the best alpha hyper parameter!! alpha 0.8!! lets plug"
      ],
      "metadata": {
        "id": "UfvXMqKZWhGN"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "# Re-train with optimal hyper parameters"
      ],
      "metadata": {
        "id": "_YbjeFmCWlwT"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "#Re train with hyper parameters\n",
        "\n",
        "features = ocr_train[train_features]\n",
        "target = ocr_train[train_target]\n",
        "\n",
        "features_test = ocr_test[test_features]\n",
        "target_test=ocr_test[test_target]\n",
        "\n",
        "#alpha = 0.8\n",
        "mnm = MultinomialNB(alpha=0.8, fit_prior=True)\n",
        "mnm.fit(features, target)\n",
        "\n",
        "prediction = mnm.predict(features_test)\n",
        "\n",
        "\n",
        "#confusion matrix\n",
        "t_features=ocr_test[test_features]\n",
        "t_target = ocr_test[test_target]\n",
        "t_target = t_target.to_numpy()\n",
        "\n",
        "\n",
        "cMatrix = confusion_matrix(t_target, prediction, labels=range(10))\n",
        "\n",
        "\n",
        "#Functions creating iterators for efficient looping - itertools\n",
        "classes = ['a','b','c','d','e','f','g','h','i','j']\n",
        "\n",
        "plt.figure(figsize=(10,10))\n",
        "\n",
        "plt.imshow(cMatrix, interpolation='nearest', cmap=plt.cm.Blues_r)\n",
        "plt.title('Confusion Matrix - Letter Classification')\n",
        "plt.colorbar()\n",
        "plt.xticks(range(10),classes)\n",
        "plt.yticks(range(10),classes)\n",
        "\n",
        "for k,v in itertools.product(range(cMatrix.shape[0]),range(cMatrix.shape[1])):\n",
        "  plt.text(v, k, cMatrix[k,v],horizontalalignment=\"center\",color=\"white\" if cMatrix[k,v]>800 else \"black\")\n",
        "\n",
        "plt.show()\n",
        "\n",
        "#New scores\n",
        "y_true = np.array(t_target)\n",
        "y_pred = np.array(prediction)\n",
        "\n",
        "\n",
        "print(classification_report(y_true, y_pred, target_names=classes))\n"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        },
        "id": "dNUIwBtIWvJX",
        "outputId": "5b034685-ea4e-4e2d-c510-d2c2b2998740"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "/usr/local/lib/python3.7/dist-packages/sklearn/base.py:493: FutureWarning: The feature names should match those that were passed during fit. Starting version 1.2, an error will be raised.\n",
            "Feature names unseen at fit time:\n",
            "- 1.17\n",
            "- 1.18\n",
            "- 1.19\n",
            "Feature names seen at fit time, yet now missing:\n",
            "- 0.108\n",
            "- 0.109\n",
            "- 0.110\n",
            "\n",
            "  warnings.warn(message, FutureWarning)\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 720x720 with 2 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "              precision    recall  f1-score   support\n",
            "\n",
            "           a       0.63      0.86      0.73      1344\n",
            "           b       0.71      0.69      0.70       428\n",
            "           c       0.73      0.64      0.68       705\n",
            "           d       0.57      0.70      0.62       481\n",
            "           e       0.80      0.62      0.70      1652\n",
            "           f       0.54      0.66      0.59       307\n",
            "           g       0.80      0.69      0.74       824\n",
            "           h       0.59      0.64      0.61       287\n",
            "           i       0.89      0.80      0.84      1638\n",
            "           j       0.28      0.48      0.36        63\n",
            "\n",
            "    accuracy                           0.72      7729\n",
            "   macro avg       0.65      0.68      0.66      7729\n",
            "weighted avg       0.74      0.72      0.72      7729\n",
            "\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "# New Model ??\n"
      ],
      "metadata": {
        "id": "k5-L5KPw4ouc"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "Hyper parameters updated yet no changes made. Let's try another model, the --ComplimentNB-- CategoricalNB. "
      ],
      "metadata": {
        "id": "5TPA0UiW4vLg"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "\n",
        "#Re train with hyper parameters\n",
        "\n",
        "features = ocr_train[train_features]\n",
        "target = ocr_train[train_target]\n",
        "\n",
        "features_test = ocr_test[test_features]\n",
        "target_test=ocr_test[test_target]\n",
        "\n",
        "#ComplimentNB model\n",
        "canb = CategoricalNB(alpha=1.0, fit_prior=True)\n",
        "canb.fit(features, target)\n",
        "\n",
        "prediction = canb.predict(features_test)\n",
        "\n",
        "\n",
        "#confusion matrix\n",
        "t_features=ocr_test[test_features]\n",
        "t_target = ocr_test[test_target]\n",
        "t_target = t_target.to_numpy()\n",
        "\n",
        "\n",
        "cMatrix = confusion_matrix(t_target, prediction, labels=range(10))\n",
        "\n",
        "\n",
        "#Functions creating iterators for efficient looping - itertools\n",
        "classes = ['a','b','c','d','e','f','g','h','i','j']\n",
        "\n",
        "plt.figure(figsize=(10,10))\n",
        "\n",
        "plt.imshow(cMatrix, interpolation='nearest', cmap=plt.cm.Reds_r)\n",
        "plt.title('Confusion Matrix - CategoricalNB Model w/ default hyperparameters')\n",
        "plt.colorbar()\n",
        "plt.xticks(range(10),classes)\n",
        "plt.yticks(range(10),classes)\n",
        "\n",
        "for k,v in itertools.product(range(cMatrix.shape[0]),range(cMatrix.shape[1])):\n",
        "  plt.text(v, k, cMatrix[k,v],horizontalalignment=\"center\",color=\"black\" if cMatrix[k,v]>800 else \"white\")\n",
        "\n",
        "plt.show()\n",
        "\n",
        "#New scores\n",
        "y_true = np.array(t_target)\n",
        "y_pred = np.array(prediction)\n",
        "\n",
        "print(\"CategoricalNB Classification Report w/ default hyperparameters\")\n",
        "print(classification_report(y_true, y_pred, target_names=classes))"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        },
        "id": "nGSbdWFd44FY",
        "outputId": "0593000f-870f-43b2-a3dd-e3ac43daf1ac"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "/usr/local/lib/python3.7/dist-packages/sklearn/base.py:493: FutureWarning: The feature names should match those that were passed during fit. Starting version 1.2, an error will be raised.\n",
            "Feature names unseen at fit time:\n",
            "- 1.17\n",
            "- 1.18\n",
            "- 1.19\n",
            "Feature names seen at fit time, yet now missing:\n",
            "- 0.108\n",
            "- 0.109\n",
            "- 0.110\n",
            "\n",
            "  warnings.warn(message, FutureWarning)\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 720x720 with 2 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "CategoricalNB Classification Report w/ default hyperparameters\n",
            "              precision    recall  f1-score   support\n",
            "\n",
            "           a       0.67      0.85      0.75      1344\n",
            "           b       0.78      0.71      0.74       428\n",
            "           c       0.75      0.72      0.73       705\n",
            "           d       0.64      0.75      0.69       481\n",
            "           e       0.82      0.64      0.72      1652\n",
            "           f       0.54      0.74      0.63       307\n",
            "           g       0.82      0.73      0.77       824\n",
            "           h       0.58      0.66      0.62       287\n",
            "           i       0.95      0.87      0.91      1638\n",
            "           j       0.30      0.56      0.39        63\n",
            "\n",
            "    accuracy                           0.76      7729\n",
            "   macro avg       0.69      0.72      0.70      7729\n",
            "weighted avg       0.78      0.76      0.76      7729\n",
            "\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "#defining parameter range\n",
        "#parameter grid takes a dictionary containing list of \n",
        "#values of each hyperparameter\n",
        "\n",
        "param_grid = {'alpha': [0.1,0.2,0.3,0.4,0.5,0.6,0.7,0.8,0.9,1.0,1.1]} \n",
        "grid = GridSearchCV(canb, param_grid, refit = True, verbose = 3)\n",
        "\n",
        "\n",
        "# fitting the model for grid search\n",
        "grid.fit(features, target)\n",
        "\n",
        "\n",
        "# print best parameter after tuning\n",
        "print(grid.best_params_)\n",
        "\n",
        "\n",
        "# print how our model looks after hyper-parameter tuning\n",
        "print(grid.best_estimator_)\n",
        "\n",
        "\n",
        "#testing the model\n",
        "grid_predictions = grid.predict(features_test)\n",
        "print(classification_report(target_test, grid_predictions))"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "sGEce0qT79zY",
        "outputId": "d92992c9-b650-49e8-df49-0af7a99b8d47"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Fitting 5 folds for each of 11 candidates, totalling 55 fits\n",
            "[CV 1/5] END .........................alpha=0.1;, score=0.745 total time=   0.1s\n",
            "[CV 2/5] END .........................alpha=0.1;, score=0.737 total time=   0.1s\n",
            "[CV 3/5] END .........................alpha=0.1;, score=0.767 total time=   0.1s\n",
            "[CV 4/5] END .........................alpha=0.1;, score=0.765 total time=   0.1s\n",
            "[CV 5/5] END .........................alpha=0.1;, score=0.765 total time=   0.1s\n",
            "[CV 1/5] END .........................alpha=0.2;, score=0.745 total time=   0.1s\n",
            "[CV 2/5] END .........................alpha=0.2;, score=0.737 total time=   0.1s\n",
            "[CV 3/5] END .........................alpha=0.2;, score=0.766 total time=   0.1s\n",
            "[CV 4/5] END .........................alpha=0.2;, score=0.765 total time=   0.2s\n",
            "[CV 5/5] END .........................alpha=0.2;, score=0.766 total time=   0.1s\n",
            "[CV 1/5] END .........................alpha=0.3;, score=0.745 total time=   0.1s\n",
            "[CV 2/5] END .........................alpha=0.3;, score=0.736 total time=   0.1s\n",
            "[CV 3/5] END .........................alpha=0.3;, score=0.766 total time=   0.2s\n",
            "[CV 4/5] END .........................alpha=0.3;, score=0.765 total time=   0.1s\n",
            "[CV 5/5] END .........................alpha=0.3;, score=0.765 total time=   0.2s\n",
            "[CV 1/5] END .........................alpha=0.4;, score=0.745 total time=   0.1s\n",
            "[CV 2/5] END .........................alpha=0.4;, score=0.736 total time=   0.2s\n",
            "[CV 3/5] END .........................alpha=0.4;, score=0.766 total time=   0.1s\n",
            "[CV 4/5] END .........................alpha=0.4;, score=0.764 total time=   0.1s\n",
            "[CV 5/5] END .........................alpha=0.4;, score=0.765 total time=   0.1s\n",
            "[CV 1/5] END .........................alpha=0.5;, score=0.745 total time=   0.1s\n",
            "[CV 2/5] END .........................alpha=0.5;, score=0.736 total time=   0.1s\n",
            "[CV 3/5] END .........................alpha=0.5;, score=0.766 total time=   0.1s\n",
            "[CV 4/5] END .........................alpha=0.5;, score=0.764 total time=   0.1s\n",
            "[CV 5/5] END .........................alpha=0.5;, score=0.765 total time=   0.1s\n",
            "[CV 1/5] END .........................alpha=0.6;, score=0.745 total time=   0.1s\n",
            "[CV 2/5] END .........................alpha=0.6;, score=0.736 total time=   0.1s\n",
            "[CV 3/5] END .........................alpha=0.6;, score=0.766 total time=   0.1s\n",
            "[CV 4/5] END .........................alpha=0.6;, score=0.764 total time=   0.1s\n",
            "[CV 5/5] END .........................alpha=0.6;, score=0.765 total time=   0.1s\n",
            "[CV 1/5] END .........................alpha=0.7;, score=0.745 total time=   0.1s\n",
            "[CV 2/5] END .........................alpha=0.7;, score=0.736 total time=   0.1s\n",
            "[CV 3/5] END .........................alpha=0.7;, score=0.766 total time=   0.1s\n",
            "[CV 4/5] END .........................alpha=0.7;, score=0.764 total time=   0.1s\n",
            "[CV 5/5] END .........................alpha=0.7;, score=0.765 total time=   0.1s\n",
            "[CV 1/5] END .........................alpha=0.8;, score=0.745 total time=   0.1s\n",
            "[CV 2/5] END .........................alpha=0.8;, score=0.736 total time=   0.1s\n",
            "[CV 3/5] END .........................alpha=0.8;, score=0.766 total time=   0.1s\n",
            "[CV 4/5] END .........................alpha=0.8;, score=0.764 total time=   0.1s\n",
            "[CV 5/5] END .........................alpha=0.8;, score=0.765 total time=   0.1s\n",
            "[CV 1/5] END .........................alpha=0.9;, score=0.745 total time=   0.1s\n",
            "[CV 2/5] END .........................alpha=0.9;, score=0.737 total time=   0.1s\n",
            "[CV 3/5] END .........................alpha=0.9;, score=0.767 total time=   0.1s\n",
            "[CV 4/5] END .........................alpha=0.9;, score=0.764 total time=   0.1s\n",
            "[CV 5/5] END .........................alpha=0.9;, score=0.765 total time=   0.1s\n",
            "[CV 1/5] END .........................alpha=1.0;, score=0.745 total time=   0.1s\n",
            "[CV 2/5] END .........................alpha=1.0;, score=0.737 total time=   0.1s\n",
            "[CV 3/5] END .........................alpha=1.0;, score=0.767 total time=   0.1s\n",
            "[CV 4/5] END .........................alpha=1.0;, score=0.764 total time=   0.1s\n",
            "[CV 5/5] END .........................alpha=1.0;, score=0.765 total time=   0.1s\n",
            "[CV 1/5] END .........................alpha=1.1;, score=0.745 total time=   0.1s\n",
            "[CV 2/5] END .........................alpha=1.1;, score=0.737 total time=   0.1s\n",
            "[CV 3/5] END .........................alpha=1.1;, score=0.766 total time=   0.1s\n",
            "[CV 4/5] END .........................alpha=1.1;, score=0.764 total time=   0.1s\n",
            "[CV 5/5] END .........................alpha=1.1;, score=0.765 total time=   0.1s\n",
            "{'alpha': 0.1}\n",
            "CategoricalNB(alpha=0.1)\n",
            "              precision    recall  f1-score   support\n",
            "\n",
            "           0       0.67      0.85      0.75      1344\n",
            "           1       0.78      0.71      0.75       428\n",
            "           2       0.75      0.72      0.73       705\n",
            "           3       0.64      0.75      0.69       481\n",
            "           4       0.82      0.64      0.72      1652\n",
            "           5       0.55      0.74      0.63       307\n",
            "           6       0.81      0.73      0.77       824\n",
            "           7       0.58      0.67      0.62       287\n",
            "           8       0.95      0.87      0.91      1638\n",
            "           9       0.32      0.56      0.41        63\n",
            "\n",
            "    accuracy                           0.76      7729\n",
            "   macro avg       0.69      0.72      0.70      7729\n",
            "weighted avg       0.78      0.76      0.76      7729\n",
            "\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "/usr/local/lib/python3.7/dist-packages/sklearn/base.py:493: FutureWarning: The feature names should match those that were passed during fit. Starting version 1.2, an error will be raised.\n",
            "Feature names unseen at fit time:\n",
            "- 1.17\n",
            "- 1.18\n",
            "- 1.19\n",
            "Feature names seen at fit time, yet now missing:\n",
            "- 0.108\n",
            "- 0.109\n",
            "- 0.110\n",
            "\n",
            "  warnings.warn(message, FutureWarning)\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "# New model new hyper parameters... right?"
      ],
      "metadata": {
        "id": "MxpTmZPZ8jm1"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "\n",
        "#Re train with hyper parameters\n",
        "\n",
        "features = ocr_train[train_features]\n",
        "target = ocr_train[train_target]\n",
        "\n",
        "features_test = ocr_test[test_features]\n",
        "target_test=ocr_test[test_target]\n",
        "\n",
        "#alpha = 0.1\n",
        "canb = CategoricalNB(alpha=0.1, fit_prior=True)\n",
        "canb.fit(features, target)\n",
        "\n",
        "prediction = canb.predict(features_test)\n",
        "\n",
        "\n",
        "#confusion matrix\n",
        "t_features=ocr_test[test_features]\n",
        "t_target = ocr_test[test_target]\n",
        "t_target = t_target.to_numpy()\n",
        "\n",
        "\n",
        "cMatrix = confusion_matrix(t_target, prediction, labels=range(10))\n",
        "\n",
        "\n",
        "#Functions creating iterators for efficient looping - itertools\n",
        "classes = ['a','b','c','d','e','f','g','h','i','j']\n",
        "\n",
        "plt.figure(figsize=(10,10))\n",
        "\n",
        "plt.imshow(cMatrix, interpolation='nearest', cmap=plt.cm.Reds_r)\n",
        "plt.title('Confusion Matrix - CategoricalNB Model w/ optimized hyperparameters')\n",
        "plt.colorbar()\n",
        "plt.xticks(range(10),classes)\n",
        "plt.yticks(range(10),classes)\n",
        "\n",
        "for k,v in itertools.product(range(cMatrix.shape[0]),range(cMatrix.shape[1])):\n",
        "  plt.text(v, k, cMatrix[k,v],horizontalalignment=\"center\",color=\"black\" if cMatrix[k,v]>800 else \"white\")\n",
        "\n",
        "plt.show()\n",
        "\n",
        "#New scores\n",
        "y_true = np.array(t_target)\n",
        "y_pred = np.array(prediction)\n",
        "\n",
        "print(\"CategoricalNB Classification Report w/ optimized hyperparameters\")\n",
        "print(classification_report(y_true, y_pred, target_names=classes))"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        },
        "id": "yTX0_gTV9EQA",
        "outputId": "91788e85-efb7-4982-bf3a-893704fe1d28"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "/usr/local/lib/python3.7/dist-packages/sklearn/base.py:493: FutureWarning: The feature names should match those that were passed during fit. Starting version 1.2, an error will be raised.\n",
            "Feature names unseen at fit time:\n",
            "- 1.17\n",
            "- 1.18\n",
            "- 1.19\n",
            "Feature names seen at fit time, yet now missing:\n",
            "- 0.108\n",
            "- 0.109\n",
            "- 0.110\n",
            "\n",
            "  warnings.warn(message, FutureWarning)\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 720x720 with 2 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "CategoricalNB Classification Report w/ optimized hyperparameters\n",
            "              precision    recall  f1-score   support\n",
            "\n",
            "           a       0.67      0.85      0.75      1344\n",
            "           b       0.78      0.71      0.75       428\n",
            "           c       0.75      0.72      0.73       705\n",
            "           d       0.64      0.75      0.69       481\n",
            "           e       0.82      0.64      0.72      1652\n",
            "           f       0.55      0.74      0.63       307\n",
            "           g       0.81      0.73      0.77       824\n",
            "           h       0.58      0.67      0.62       287\n",
            "           i       0.95      0.87      0.91      1638\n",
            "           j       0.32      0.56      0.41        63\n",
            "\n",
            "    accuracy                           0.76      7729\n",
            "   macro avg       0.69      0.72      0.70      7729\n",
            "weighted avg       0.78      0.76      0.76      7729\n",
            "\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "ROC New"
      ],
      "metadata": {
        "id": "Jr6MKPZLFIAr"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "\n",
        "\n",
        "lw = 2\n",
        "plt.figure()\n",
        "\n",
        "\n",
        "#range represents classes a-j\n",
        "for rocClass in range(10):\n",
        "  predict_prob = mnm.predict_proba(features_test)\n",
        "  prob_possibility = predict_prob[:, rocClass]\n",
        "  thresholds = np.arange(0.0, 1.2, 0.1) #just as range gives a range of numbers and its skip patter so does numpy arange\n",
        "  true_pos, false_pos = [0] * len(thresholds), [0] * len(thresholds)\n",
        "  for pred, y in zip(prob_possibility, target_test):\n",
        "    for i, threshold in enumerate(thresholds):\n",
        "      if pred >= threshold:\n",
        "        if y == rocClass:\n",
        "          true_pos[i] += 1\n",
        "        else:\n",
        "          false_pos[i] += 1\n",
        "      else:\n",
        "        break\n",
        "  true_pos_rate = [tp/len(np.where(target_test == rocClass)[0]) for tp in true_pos]\n",
        "  false_pos_rate = [fp/len(np.where(target_test != rocClass)[0]) for fp in false_pos]\n",
        "  roc_score = roc_auc_score(target_test == rocClass, prob_possibility, max_fpr=1)\n",
        "  plt.plot(false_pos_rate, true_pos_rate, label='ROC curve of class {0} (area = {1:0.2f})'.format(classes[rocClass], roc_score))\n",
        "  plt.legend(loc=\"lower right\")\n",
        "  plt.xlim([0.0,1.0])\n",
        "plt.ylim([0.0, 1.0])\n",
        "plt.xlabel('False Positive Rate')\n"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        },
        "id": "7o5q8Z2qFHSC",
        "outputId": "aba09fd6-be13-4d80-e81c-15d8622b9eeb"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "/usr/local/lib/python3.7/dist-packages/sklearn/base.py:493: FutureWarning: The feature names should match those that were passed during fit. Starting version 1.2, an error will be raised.\n",
            "Feature names unseen at fit time:\n",
            "- 1.17\n",
            "- 1.18\n",
            "- 1.19\n",
            "Feature names seen at fit time, yet now missing:\n",
            "- 0.108\n",
            "- 0.109\n",
            "- 0.110\n",
            "\n",
            "  warnings.warn(message, FutureWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/base.py:493: FutureWarning: The feature names should match those that were passed during fit. Starting version 1.2, an error will be raised.\n",
            "Feature names unseen at fit time:\n",
            "- 1.17\n",
            "- 1.18\n",
            "- 1.19\n",
            "Feature names seen at fit time, yet now missing:\n",
            "- 0.108\n",
            "- 0.109\n",
            "- 0.110\n",
            "\n",
            "  warnings.warn(message, FutureWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/base.py:493: FutureWarning: The feature names should match those that were passed during fit. Starting version 1.2, an error will be raised.\n",
            "Feature names unseen at fit time:\n",
            "- 1.17\n",
            "- 1.18\n",
            "- 1.19\n",
            "Feature names seen at fit time, yet now missing:\n",
            "- 0.108\n",
            "- 0.109\n",
            "- 0.110\n",
            "\n",
            "  warnings.warn(message, FutureWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/base.py:493: FutureWarning: The feature names should match those that were passed during fit. Starting version 1.2, an error will be raised.\n",
            "Feature names unseen at fit time:\n",
            "- 1.17\n",
            "- 1.18\n",
            "- 1.19\n",
            "Feature names seen at fit time, yet now missing:\n",
            "- 0.108\n",
            "- 0.109\n",
            "- 0.110\n",
            "\n",
            "  warnings.warn(message, FutureWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/base.py:493: FutureWarning: The feature names should match those that were passed during fit. Starting version 1.2, an error will be raised.\n",
            "Feature names unseen at fit time:\n",
            "- 1.17\n",
            "- 1.18\n",
            "- 1.19\n",
            "Feature names seen at fit time, yet now missing:\n",
            "- 0.108\n",
            "- 0.109\n",
            "- 0.110\n",
            "\n",
            "  warnings.warn(message, FutureWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/base.py:493: FutureWarning: The feature names should match those that were passed during fit. Starting version 1.2, an error will be raised.\n",
            "Feature names unseen at fit time:\n",
            "- 1.17\n",
            "- 1.18\n",
            "- 1.19\n",
            "Feature names seen at fit time, yet now missing:\n",
            "- 0.108\n",
            "- 0.109\n",
            "- 0.110\n",
            "\n",
            "  warnings.warn(message, FutureWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/base.py:493: FutureWarning: The feature names should match those that were passed during fit. Starting version 1.2, an error will be raised.\n",
            "Feature names unseen at fit time:\n",
            "- 1.17\n",
            "- 1.18\n",
            "- 1.19\n",
            "Feature names seen at fit time, yet now missing:\n",
            "- 0.108\n",
            "- 0.109\n",
            "- 0.110\n",
            "\n",
            "  warnings.warn(message, FutureWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/base.py:493: FutureWarning: The feature names should match those that were passed during fit. Starting version 1.2, an error will be raised.\n",
            "Feature names unseen at fit time:\n",
            "- 1.17\n",
            "- 1.18\n",
            "- 1.19\n",
            "Feature names seen at fit time, yet now missing:\n",
            "- 0.108\n",
            "- 0.109\n",
            "- 0.110\n",
            "\n",
            "  warnings.warn(message, FutureWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/base.py:493: FutureWarning: The feature names should match those that were passed during fit. Starting version 1.2, an error will be raised.\n",
            "Feature names unseen at fit time:\n",
            "- 1.17\n",
            "- 1.18\n",
            "- 1.19\n",
            "Feature names seen at fit time, yet now missing:\n",
            "- 0.108\n",
            "- 0.109\n",
            "- 0.110\n",
            "\n",
            "  warnings.warn(message, FutureWarning)\n",
            "/usr/local/lib/python3.7/dist-packages/sklearn/base.py:493: FutureWarning: The feature names should match those that were passed during fit. Starting version 1.2, an error will be raised.\n",
            "Feature names unseen at fit time:\n",
            "- 1.17\n",
            "- 1.18\n",
            "- 1.19\n",
            "Feature names seen at fit time, yet now missing:\n",
            "- 0.108\n",
            "- 0.109\n",
            "- 0.110\n",
            "\n",
            "  warnings.warn(message, FutureWarning)\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "Text(0.5, 0, 'False Positive Rate')"
            ]
          },
          "metadata": {},
          "execution_count": 16
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "In conclusion, as a comparrison, the models f1-score performance is as follows:\n",
        "1. **CategoricalNB**\n",
        "2. **MultiNominalNB**\n",
        "3. **ComplimentaryNB**\n",
        "\n",
        "By combining a new classifier that's suitable for classification with discrete features, together with optimized hyperparameters we were able to achieve a result higher than the example given. We have reached our goal and have successfully raised the score of every classification including the lowest classification 'j'. "
      ],
      "metadata": {
        "id": "fc8adVcFCRcS"
      }
    },
    {
      "cell_type": "code",
      "source": [
        ""
      ],
      "metadata": {
        "id": "Oci4TuR7FGxW"
      },
      "execution_count": null,
      "outputs": []
    }
  ]
}