Digit_recognizer.ipynb

{
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "6pfyVThcLHrl"
      },
      "source": [
        "## Digit Recognizer\n",
        "https://www.kaggle.com/c/digit-recognizer"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "collapsed": true,
        "id": "AjJIW2nXLHrp"
      },
      "outputs": [],
      "source": [
        "%matplotlib inline\n",
        "import cv2\n",
        "import numpy as np\n",
        "import matplotlib.pyplot as plt\n",
        "import pandas as pd\n"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "xugkbcKkLHrr"
      },
      "source": [
        "## Загружаем исходные данные"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "train = np.loadtxt('train.csv', delimiter=',', skiprows=1)\n",
        "test = np.loadtxt('test.csv', delimiter=',', skiprows=1)"
      ],
      "metadata": {
        "id": "Wt0vCL6StxXK"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "collapsed": true,
        "id": "iTIugXUcLHrs"
      },
      "outputs": [],
      "source": [
        "# сохраняем разметку в отдельную переменную\n",
        "train_label = train[:, 0]\n",
        "# приводим размерность к удобному для обаботки виду\n",
        "train_img = np.resize(train[:, 1:], (train.shape[0], 28, 28))\n",
        "test_img = np.resize(test, (test.shape[0], 28, 28))"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "Ut1XZHyrLHrs",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "0cb30f16-499f-464b-9912-eb62e6779514"
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(42000, 28, 28)"
            ]
          },
          "metadata": {},
          "execution_count": 4
        }
      ],
      "source": [
        "train_img.shape"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "WkMo1QwyLHru"
      },
      "source": [
        "## Визуализируем исходные данные"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "VMgigpyHLHru",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 272
        },
        "outputId": "50e238db-0541-441a-e9fd-1cac7cdcb74c"
      },
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 2000x1000 with 5 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        }
      ],
      "source": [
        "fig = plt.figure(figsize=(20, 10))\n",
        "for i, img in enumerate(train_img[0:5], 1):\n",
        "    subplot = fig.add_subplot(1, 7, i)\n",
        "    plt.imshow(img, cmap='gray');\n",
        "    subplot.set_title('%s' % train_label[i - 1]);"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "VcERALp2LHrv"
      },
      "source": [
        "## Вычисляем X и Y составляющие градиента с помощью оператора Собеля"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "collapsed": true,
        "id": "Yg85zrUdLHrv"
      },
      "outputs": [],
      "source": [
        "train_sobel_x = np.zeros_like(train_img)\n",
        "train_sobel_y = np.zeros_like(train_img)\n",
        "for i in range(len(train_img)):\n",
        "    train_sobel_x[i] = cv2.Sobel(train_img[i], cv2.CV_64F, dx=1, dy=0, ksize=3)\n",
        "    train_sobel_y[i] = cv2.Sobel(train_img[i], cv2.CV_64F, dx=0, dy=1, ksize=3)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "collapsed": true,
        "id": "9DoQYxyELHrv"
      },
      "outputs": [],
      "source": [
        "test_sobel_x = np.zeros_like(test_img)\n",
        "test_sobel_y = np.zeros_like(test_img)\n",
        "for i in range(len(test_img)):\n",
        "    test_sobel_x[i] = cv2.Sobel(test_img[i], cv2.CV_64F, dx=1, dy=0, ksize=3)\n",
        "    test_sobel_y[i] = cv2.Sobel(test_img[i], cv2.CV_64F, dx=0, dy=1, ksize=3)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "ASwd5326LHrw"
      },
      "source": [
        "## Вычисляем угол и длину вектора градиента"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "collapsed": true,
        "id": "AvkqWureLHrw"
      },
      "outputs": [],
      "source": [
        "train_g, train_theta = cv2.cartToPolar(train_sobel_x, train_sobel_y)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "collapsed": true,
        "id": "q17lZls3LHrx"
      },
      "outputs": [],
      "source": [
        "test_g, test_theta = cv2.cartToPolar(test_sobel_x, test_sobel_y)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "aCZlhC7hLHrx",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 540
        },
        "outputId": "5eb07ccd-00dc-49e7-99a4-6fc0a6c10054"
      },
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 2000x1000 with 10 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        }
      ],
      "source": [
        "fig = plt.figure(figsize=(20, 10))\n",
        "for i, img in enumerate(train_g[:5], 1):\n",
        "    subplot = fig.add_subplot(1, 7, i)\n",
        "    plt.imshow(img, cmap='gray');\n",
        "    subplot.set_title('%s' % train_label[i - 1]);\n",
        "    subplot = fig.add_subplot(3, 7, i)\n",
        "    plt.hist(train_theta[i - 1].flatten(),\n",
        "             bins=16, weights=train_g[i - 1].flatten())"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "RnEzoKsJLHrx"
      },
      "source": [
        "## Вычисляем гистограммы градиентов"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "collapsed": true,
        "id": "DXg-JKiALHry"
      },
      "outputs": [],
      "source": [
        "# Гистограммы вычисляются с учетом длины вектора градиента\n",
        "train_hist = np.zeros((len(train_img), 16))\n",
        "for i in range(len(train_img)):\n",
        "    hist, borders = np.histogram(train_theta[i],\n",
        "                                 bins=16,\n",
        "                                 range=(0., 2. * np.pi),\n",
        "                                 weights=train_g[i])\n",
        "    train_hist[i] = hist"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "collapsed": true,
        "id": "yquvVxdkLHry"
      },
      "outputs": [],
      "source": [
        "test_hist = np.zeros((len(test_img), 16))\n",
        "for i in range(len(test_img)):\n",
        "    hist, borders = np.histogram(test_theta[i],\n",
        "                                 bins=16,\n",
        "                                 range=(0., 2. * np.pi),\n",
        "                                 weights=test_g[i])\n",
        "    test_hist[i] = hist"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "9Xh7cg1dLHry"
      },
      "source": [
        "## Нормируем вектор гистограммы"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "collapsed": true,
        "id": "EH2QwdEDLHrz"
      },
      "outputs": [],
      "source": [
        "# По умолчанию используется L2 норма\n",
        "train_hist = train_hist / np.linalg.norm(train_hist, axis=1)[:, None]"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "collapsed": true,
        "id": "tSix4MC-LHrz"
      },
      "outputs": [],
      "source": [
        "test_hist = test_hist / np.linalg.norm(test_hist, axis=1)[:, None]"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "66JFRaLpLHrz"
      },
      "source": [
        "## Разбиваем выборку на обучение и валидацию"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "collapsed": true,
        "id": "T2YQg5dxLHrz"
      },
      "outputs": [],
      "source": [
        "from sklearn.model_selection import train_test_split\n",
        "y_train, y_val, x_train, x_val = train_test_split(\n",
        "    train_label, train_hist, test_size=0.2,\n",
        "    random_state=42)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "J7bh3mYALHr0"
      },
      "source": [
        "## Вычисляем центроиды каждого из 10 классов"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "collapsed": true,
        "id": "Iett9cxkLHr0"
      },
      "outputs": [],
      "source": [
        "centroids = np.zeros((10, train_hist.shape[1]), dtype=np.float32)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "collapsed": true,
        "id": "c9azu8NWLHr0"
      },
      "outputs": [],
      "source": [
        "for i in range(10):\n",
        "    centroids[i] = np.mean(x_train[y_train == i], axis=0)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "zVeOxpWkLHr0"
      },
      "source": [
        "## Находим ближайший центройд для валидации"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "collapsed": true,
        "id": "kn-i_cJQLHr1"
      },
      "outputs": [],
      "source": [
        "pred_val = np.zeros_like(y_val)\n",
        "for i in range(len(y_val)):\n",
        "    distances = np.linalg.norm(centroids - x_val[i], axis=1)\n",
        "    pred_val[i] = np.argmin(distances)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "fapATONMLHr1"
      },
      "source": [
        "## Оцениваем качество решение на валидационной выборке"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "QfAtb9dALHr1",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "5076e178-8222-4e63-9963-0040a39f0fd6"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Accuracy: 0.4270238095238095\n"
          ]
        }
      ],
      "source": [
        "from sklearn.metrics import accuracy_score\n",
        "print('Accuracy: %s' % accuracy_score(y_val, pred_val))"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "jEKekatvLHr1",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "22fd0f26-21e3-4538-92de-18f6564d2593"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "              precision    recall  f1-score   support\n",
            "\n",
            "         0.0       0.34      0.12      0.18       816\n",
            "         1.0       0.59      0.76      0.66       909\n",
            "         2.0       0.33      0.29      0.31       846\n",
            "         3.0       0.55      0.49      0.51       937\n",
            "         4.0       0.33      0.31      0.32       839\n",
            "         5.0       0.46      0.67      0.55       702\n",
            "         6.0       0.24      0.45      0.31       785\n",
            "         7.0       0.73      0.41      0.53       893\n",
            "         8.0       0.39      0.49      0.44       835\n",
            "         9.0       0.48      0.27      0.34       838\n",
            "\n",
            "    accuracy                           0.43      8400\n",
            "   macro avg       0.44      0.43      0.42      8400\n",
            "weighted avg       0.45      0.43      0.42      8400\n",
            "\n"
          ]
        }
      ],
      "source": [
        "from sklearn.metrics import classification_report\n",
        "print(classification_report(y_val, pred_val))"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "cWm_E3zsLHr1",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "43903100-4631-4553-e007-57aade8fb3f6"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[[102  24  82  30 106  18 221   7 179  47]\n",
            " [  0 692   2   0  44   1 139  17  13   1]\n",
            " [ 21   0 245 158  12 203  54  16 119  18]\n",
            " [ 14   0 113 456   3 235   7   2  98   9]\n",
            " [ 11 248   5   2 260   0 205  43  37  28]\n",
            " [  3   0  84  97   5 472   3   4  33   1]\n",
            " [ 53  26  66  11 116   4 355  13  76  65]\n",
            " [  4 113  38  16  66  41 189 368  15  43]\n",
            " [ 81   6  68  44  36  44 107   3 411  35]\n",
            " [ 11  73  31  20 149  11 222  30  65 226]]\n"
          ]
        }
      ],
      "source": [
        "from sklearn.metrics import confusion_matrix\n",
        "print(confusion_matrix(y_val, pred_val))"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "77A_UJjdLHr2"
      },
      "source": [
        "## Предсказания на тестовых данных"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "collapsed": true,
        "id": "adxeSShcLHr2"
      },
      "outputs": [],
      "source": [
        "pred_test = np.zeros(len(test_img), np.uint8)\n",
        "for i in range(len(test_img)):\n",
        "    distances = np.linalg.norm(centroids - test_hist[i], axis=1)\n",
        "    pred_test[i] = np.argmin(distances)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "xpIemLc0LHr2"
      },
      "source": [
        "## Визуализируем предсказания"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "bSfS_m3OLHr3",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 272
        },
        "outputId": "39526f28-c0de-4ece-aa55-a04d7710e2c1"
      },
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 2000x1000 with 5 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        }
      ],
      "source": [
        "fig = plt.figure(figsize=(20, 10))\n",
        "for i, img in enumerate(test_img[0:5], 1):\n",
        "    subplot = fig.add_subplot(1, 7, i)\n",
        "    plt.imshow(img, cmap='gray');\n",
        "    subplot.set_title('%s' % pred_test[i - 1]);"
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "KNN"
      ],
      "metadata": {
        "id": "UlOgCSsh8XM0"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "from sklearn.neighbors import KNeighborsClassifier\n",
        "from sklearn.metrics import accuracy_score\n",
        "\n",
        "clf = KNeighborsClassifier()\n",
        "clf.fit(x_train, y_train)\n",
        "pred = clf.predict(x_val)\n",
        "accuracy_score(y_val, pred)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "WsVTP-6G6Xeq",
        "outputId": "9fdaccb0-7ae7-48af-8cd5-abb81458f7e3"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "0.5813095238095238"
            ]
          },
          "metadata": {},
          "execution_count": 24
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "from sklearn.model_selection import GridSearchCV\n",
        "grid_params = {\"n_neighbors\": range(4, 100),\n",
        "               \"weights\": [\"uniform\", \"distance\"]}\n",
        "knn = KNeighborsClassifier()\n",
        "grid_search = GridSearchCV(knn, grid_params, verbose=2, n_jobs=-1)\n",
        "grid_search.fit(x_val, y_val)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 137
        },
        "id": "sCL_JVDyDBor",
        "outputId": "d39186a5-0b2d-4e47-9316-53aadb770feb"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Fitting 5 folds for each of 192 candidates, totalling 960 fits\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "GridSearchCV(estimator=KNeighborsClassifier(), n_jobs=-1,\n",
              "             param_grid={'n_neighbors': range(4, 100),\n",
              "                         'weights': ['uniform', 'distance']},\n",
              "             verbose=2)"
            ],
            "text/html": [
              "<style>#sk-container-id-1 {color: black;background-color: white;}#sk-container-id-1 pre{padding: 0;}#sk-container-id-1 div.sk-toggleable {background-color: white;}#sk-container-id-1 label.sk-toggleable__label {cursor: pointer;display: block;width: 100%;margin-bottom: 0;padding: 0.3em;box-sizing: border-box;text-align: center;}#sk-container-id-1 label.sk-toggleable__label-arrow:before {content: \"▸\";float: left;margin-right: 0.25em;color: #696969;}#sk-container-id-1 label.sk-toggleable__label-arrow:hover:before {color: black;}#sk-container-id-1 div.sk-estimator:hover label.sk-toggleable__label-arrow:before {color: black;}#sk-container-id-1 div.sk-toggleable__content {max-height: 0;max-width: 0;overflow: hidden;text-align: left;background-color: #f0f8ff;}#sk-container-id-1 div.sk-toggleable__content pre {margin: 0.2em;color: black;border-radius: 0.25em;background-color: #f0f8ff;}#sk-container-id-1 input.sk-toggleable__control:checked~div.sk-toggleable__content {max-height: 200px;max-width: 100%;overflow: auto;}#sk-container-id-1 input.sk-toggleable__control:checked~label.sk-toggleable__label-arrow:before {content: \"▾\";}#sk-container-id-1 div.sk-estimator input.sk-toggleable__control:checked~label.sk-toggleable__label {background-color: #d4ebff;}#sk-container-id-1 div.sk-label input.sk-toggleable__control:checked~label.sk-toggleable__label {background-color: #d4ebff;}#sk-container-id-1 input.sk-hidden--visually {border: 0;clip: rect(1px 1px 1px 1px);clip: rect(1px, 1px, 1px, 1px);height: 1px;margin: -1px;overflow: hidden;padding: 0;position: absolute;width: 1px;}#sk-container-id-1 div.sk-estimator {font-family: monospace;background-color: #f0f8ff;border: 1px dotted black;border-radius: 0.25em;box-sizing: border-box;margin-bottom: 0.5em;}#sk-container-id-1 div.sk-estimator:hover {background-color: #d4ebff;}#sk-container-id-1 div.sk-parallel-item::after {content: \"\";width: 100%;border-bottom: 1px solid gray;flex-grow: 1;}#sk-container-id-1 div.sk-label:hover label.sk-toggleable__label {background-color: #d4ebff;}#sk-container-id-1 div.sk-serial::before {content: \"\";position: absolute;border-left: 1px solid gray;box-sizing: border-box;top: 0;bottom: 0;left: 50%;z-index: 0;}#sk-container-id-1 div.sk-serial {display: flex;flex-direction: column;align-items: center;background-color: white;padding-right: 0.2em;padding-left: 0.2em;position: relative;}#sk-container-id-1 div.sk-item {position: relative;z-index: 1;}#sk-container-id-1 div.sk-parallel {display: flex;align-items: stretch;justify-content: center;background-color: white;position: relative;}#sk-container-id-1 div.sk-item::before, #sk-container-id-1 div.sk-parallel-item::before {content: \"\";position: absolute;border-left: 1px solid gray;box-sizing: border-box;top: 0;bottom: 0;left: 50%;z-index: -1;}#sk-container-id-1 div.sk-parallel-item {display: flex;flex-direction: column;z-index: 1;position: relative;background-color: white;}#sk-container-id-1 div.sk-parallel-item:first-child::after {align-self: flex-end;width: 50%;}#sk-container-id-1 div.sk-parallel-item:last-child::after {align-self: flex-start;width: 50%;}#sk-container-id-1 div.sk-parallel-item:only-child::after {width: 0;}#sk-container-id-1 div.sk-dashed-wrapped {border: 1px dashed gray;margin: 0 0.4em 0.5em 0.4em;box-sizing: border-box;padding-bottom: 0.4em;background-color: white;}#sk-container-id-1 div.sk-label label {font-family: monospace;font-weight: bold;display: inline-block;line-height: 1.2em;}#sk-container-id-1 div.sk-label-container {text-align: center;}#sk-container-id-1 div.sk-container {/* jupyter's `normalize.less` sets `[hidden] { display: none; }` but bootstrap.min.css set `[hidden] { display: none !important; }` so we also need the `!important` here to be able to override the default hidden behavior on the sphinx rendered scikit-learn.org. See: https://github.com/scikit-learn/scikit-learn/issues/21755 */display: inline-block !important;position: relative;}#sk-container-id-1 div.sk-text-repr-fallback {display: none;}</style><div id=\"sk-container-id-1\" class=\"sk-top-container\"><div class=\"sk-text-repr-fallback\"><pre>GridSearchCV(estimator=KNeighborsClassifier(), n_jobs=-1,\n",
              "             param_grid={&#x27;n_neighbors&#x27;: range(4, 100),\n",
              "                         &#x27;weights&#x27;: [&#x27;uniform&#x27;, &#x27;distance&#x27;]},\n",
              "             verbose=2)</pre><b>In a Jupyter environment, please rerun this cell to show the HTML representation or trust the notebook. <br />On GitHub, the HTML representation is unable to render, please try loading this page with nbviewer.org.</b></div><div class=\"sk-container\" hidden><div class=\"sk-item sk-dashed-wrapped\"><div class=\"sk-label-container\"><div class=\"sk-label sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-1\" type=\"checkbox\" ><label for=\"sk-estimator-id-1\" class=\"sk-toggleable__label sk-toggleable__label-arrow\">GridSearchCV</label><div class=\"sk-toggleable__content\"><pre>GridSearchCV(estimator=KNeighborsClassifier(), n_jobs=-1,\n",
              "             param_grid={&#x27;n_neighbors&#x27;: range(4, 100),\n",
              "                         &#x27;weights&#x27;: [&#x27;uniform&#x27;, &#x27;distance&#x27;]},\n",
              "             verbose=2)</pre></div></div></div><div class=\"sk-parallel\"><div class=\"sk-parallel-item\"><div class=\"sk-item\"><div class=\"sk-label-container\"><div class=\"sk-label sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-2\" type=\"checkbox\" ><label for=\"sk-estimator-id-2\" class=\"sk-toggleable__label sk-toggleable__label-arrow\">estimator: KNeighborsClassifier</label><div class=\"sk-toggleable__content\"><pre>KNeighborsClassifier()</pre></div></div></div><div class=\"sk-serial\"><div class=\"sk-item\"><div class=\"sk-estimator sk-toggleable\"><input class=\"sk-toggleable__control sk-hidden--visually\" id=\"sk-estimator-id-3\" type=\"checkbox\" ><label for=\"sk-estimator-id-3\" class=\"sk-toggleable__label sk-toggleable__label-arrow\">KNeighborsClassifier</label><div class=\"sk-toggleable__content\"><pre>KNeighborsClassifier()</pre></div></div></div></div></div></div></div></div></div></div>"
            ]
          },
          "metadata": {},
          "execution_count": 25
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "grid_search.best_params_"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "wpNQ_GndElmm",
        "outputId": "dea91cbf-724b-4338-c148-8e6460425986"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "{'n_neighbors': 20, 'weights': 'distance'}"
            ]
          },
          "metadata": {},
          "execution_count": 26
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "clf = KNeighborsClassifier(n_neighbors=20, weights='distance')\n",
        "clf.fit(x_train, y_train)\n",
        "pred = clf.predict(x_val)\n",
        "accuracy_score(y_val, pred)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "qmgXBgslEliN",
        "outputId": "83157dc5-f3ae-4ec1-cc90-3e58c0dda1ff"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "0.6202380952380953"
            ]
          },
          "metadata": {},
          "execution_count": 27
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "pred = clf.predict(test_hist)"
      ],
      "metadata": {
        "id": "ZFghpCaC9GJR"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "img = pred.astype(np.uint8)"
      ],
      "metadata": {
        "id": "J9G0BoFRPNcq"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "AqfqkzHmLHr3"
      },
      "source": [
        "## Готовим файл для отправки"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "with open('submit.txt', 'w') as dst:\n",
        "    dst.write('ImageId,Label\\n')\n",
        "    for i, p in enumerate(img, 1):\n",
        "        dst.write('%s,%s\\n' % (i, p))"
      ],
      "metadata": {
        "id": "DSq9IgStGX22"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "!head submit.txt"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "XQvGd-aKLoje",
        "outputId": "452096e8-396a-4027-e57c-2024b3522bec"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "ImageId,Label\n",
            "1,3\n",
            "2,5\n",
            "3,9\n",
            "4,0\n",
            "5,3\n",
            "6,7\n",
            "7,0\n",
            "8,3\n",
            "9,3\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "collapsed": true,
        "id": "0z3ZpgzULHr-"
      },
      "outputs": [],
      "source": [
        "##Your submission scored 0.61878"
      ]
    }
  ],
  "metadata": {
    "kernelspec": {
      "display_name": "Python 3",
      "language": "python",
      "name": "python3"
    },
    "language_info": {
      "codemirror_mode": {
        "name": "ipython",
        "version": 3
      },
      "file_extension": ".py",
      "mimetype": "text/x-python",
      "name": "python",
      "nbconvert_exporter": "python",
      "pygments_lexer": "ipython3",
      "version": "3.6.4"
    },
    "colab": {
      "provenance": []
    }
  },
  "nbformat": 4,
  "nbformat_minor": 0
}