Task1.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "Meshcheriakov_Ilya_Task1",
      "provenance": [],
      "collapsed_sections": [],
      "toc_visible": true
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    }
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "JKKjEkgRpla1"
      },
      "source": [
        "# Релизуйте свой классификатор KNN"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "3-P3mD4h44Ga"
      },
      "source": [
        "### Импорты"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "53zgksKPu752"
      },
      "source": [
        "import numpy as np\n",
        "import pandas as pd\n",
        "from sklearn.datasets import make_moons\n",
        "from sklearn.neighbors import KNeighborsClassifier\n",
        "from sklearn.model_selection import train_test_split\n",
        "import matplotlib.pyplot as plt\n",
        "from sklearn.metrics import accuracy_score\n",
        "%matplotlib inline"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "lCPWZO494-VI"
      },
      "source": [
        "### Реализуйте класс MyKNeighborsClassifier (6 баллов)\n",
        "\n",
        "1. (3 балла) Реализован класс с параметрами по умолчанию\n",
        "2. (1 балл) Реализовано манхетенское расстояние `(metric='manhattan')`\n",
        "3. (2 балла) Реализованы веса, зависящие от растояния `(weights='distance')`\n",
        "\n"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "6t8ST3JapfKE"
      },
      "source": [
        "class MyKNeighborsClassifier:\n",
        "  def __init__(self, n_neighbours: int = 5, weights: str = \"uniform\", metric: str = \"euclidean\"):\n",
        "    if metric not in [\"manhattan\", \"euclidean\"]:\n",
        "      raise ValueError(\"This type of metric is not supported: {}\".format(metric))\n",
        "    if weights not in [\"uniform\", \"distance\"]:\n",
        "      raise ValueError(\"This type of weights is not supported: {}\".format(weights))\n",
        "    if n_neighbours == 0: # не даем пользователю ввести 0 соседей\n",
        "      raise ValueError(\"Expected n_neighbors > 0. Got 0\")\n",
        "    \n",
        "    self.n_neighbours = n_neighbours\n",
        "    self.weights = weights\n",
        "    self.metric = metric\n",
        "\n",
        "    self.X = None\n",
        "    self.y = None\n",
        "\n",
        "  def fit(self, X: np.ndarray, y: np.ndarray) -> None:\n",
        "    self.train_set = X\n",
        "    self.train_class = y\n",
        "\n",
        "  def predict(self, X: np.ndarray) -> np.ndarray:\n",
        "    result_list = [] # список, в который будут заноситься предсказанные моделью классы\n",
        "    for point in X: # для каждой точки тестовой выборки делаем следующее:\n",
        "      if self.metric == 'manhattan':\n",
        "        testData = [[self.dist_m(point, self.train_set[i]), int(self.train_class[i])] for i in range(len(self.train_set))] # Создаем список формата [euclidean_distance, class] для каждой точки тренировочной выборки\n",
        "      elif self.metric == 'euclidean':\n",
        "        testData = [[self.dist_e(point, self.train_set[i]), int(self.train_class[i])] for i in range(len(self.train_set))] # Создаем список формата [manhattan_distance, class] для каждой точки тренировочной выборки\n",
        "      class_counter = np.zeros(len(np.unique(self.train_class)), dtype='float') # создаем массив из нулей, размер которого совпадает с количеством классов\n",
        "      if self.weights == 'uniform':\n",
        "        for elem in sorted(testData)[0:self.n_neighbours]:\n",
        "          class_counter[elem[1]] += 1 # инкрементируем элемент массива, порядковый номер которого совпадает с классом очередного \"соседа\"\n",
        "      elif self.weights == 'distance':\n",
        "        for elem in sorted(testData)[0:self.n_neighbours]:\n",
        "          class_counter[elem[1]] += 1/elem[0] # записываем вес \"соседа\" в элемент массива, порядковый номер которого совпадает с классом этого очередного \"соседа\"\n",
        "      result = sorted(zip(class_counter, range(len(np.unique(self.train_class)))), reverse=True) # кортеж формата [(weignt_class_0, class_0), (weight_class_1, class_1), ...], отсортированный по убыванию weight_class\n",
        "      if result[0][0] == result[1][0]: # при одинаковых максимальных весах (количестве) выбирается тот класс, порядковый номер которого меньше.\n",
        "        result_list.append(result[1][1])\n",
        "      else:\n",
        "        result_list.append(result[0][1])\n",
        "      result_array = np.ndarray(shape=(len(result_list), 1), dtype=int, buffer=np.array(result_list)) # создаем массив из списка\n",
        "    return result_array\n",
        "\n",
        "  def dist_e(self, a, b): # вычисление евклидова расстояния\n",
        "    return ((a[0] - b[0]) ** 2 + (a[1] - b[1]) ** 2) ** 0.5\n",
        "  \n",
        "  def dist_m(self, a, b): # вычисление манхэттенского расстояния\n",
        "    return abs(a[0] - b[0]) + abs(a[1] - b[1])"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "qCTuvZTkPw-v",
        "outputId": "4db08b3b-2bb6-483c-f822-8ec0d191cf9a"
      },
      "source": [
        "### Тестовый датасет на проверку работоспособности модели\r\n",
        "X_train_start = np.ndarray(shape=(5, 2), dtype=int, buffer=np.array([[1, 1], [1, 2], [1, 3], [1, 4], [1, 5]]))\r\n",
        "y_train_start = np.ndarray(shape=(5, 1), dtype=int, buffer=np.array([0, 1, 1, 0, 1]))\r\n",
        "T = np.ndarray(shape=(2, 2), dtype=int, buffer=np.array([[2, 3], [3, 3]]))\r\n",
        "\r\n",
        "knn_start = MyKNeighborsClassifier(n_neighbours=3, weights=\"uniform\", metric=\"euclidean\")\r\n",
        "knn_start.fit(X_train_start, y_train_start)\r\n",
        "y_pred_start = knn_start.predict(T)\r\n",
        "y_pred_start"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array([[1],\n",
              "       [1]])"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 172
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "ZTXberyKSzjX"
      },
      "source": [
        "def MyAccuracyScore (y_true, y_pred): # описание своего вычислителя доли правильных ответов\r\n",
        "  return sum([int(y_pred[i]==y_true[i]) for i in range(len(y_true))]) / float(len(y_true))"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "WI_Tq_3W05aY"
      },
      "source": [
        "### Проверьте решение на датасете и сравните с kNeighborsClassifier из sklearn (4 балла)\n",
        "1. (1 балл) Выведите accuracy_score для вашего решения и для решения из sklearn, посчитайте в 4 вариациях (Должно получиться 8 чисел)\n",
        "  1. Параметры по умолчанию\n",
        "  2. `weights='distance'`\n",
        "  3. `metric='manhattan'`\n",
        "  4. `weights='distance'`, `metric='manhattan'`\n",
        "2. (2 балла) Переберите параметр k от 1 до 10 для каждой модели из пункта выше (получится 4 графика по две линии на каждом)\n",
        "  1. Переберите параметр, каждый раз обучайте модель\n",
        "  2. Выведите график зависимости `accuracy` от `k`\n",
        "  3. На этом же графике выведите пунктирной линией такую же зависимость для модели из sklearn\n",
        "3. (3 балла) Напишите вывод, сравнение всего, что получилось – получилось ли у вас достичь таких же результатов, как в sklearn, как на результат влияют параметры, какая модель и с какими параметрами оказалась лучшей.\n",
        "\n"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 282
        },
        "id": "3HdqVBkH1SQj",
        "outputId": "3f14ef67-1f4c-418c-8a14-796102e017ce"
      },
      "source": [
        "X, y = make_moons(n_samples=500, noise=0.3, random_state=42)\n",
        "plt.scatter(X[:, 0], X[:, 1], c=y)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<matplotlib.collections.PathCollection at 0x7f9d43050790>"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 174
        },
        {
          "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": {
        "id": "a-vtohd34EpK"
      },
      "source": [
        "X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2, random_state=42)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "govUWI-257_Z"
      },
      "source": [
        "1. Выведите accuracy_score для вашего решения и для решения из sklearn, посчитайте в 4 вариациях (Должно получиться 8 чисел)"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "8ipTQXS2_7J8",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "f57af03f-87e6-4e16-e1b0-c5eb11ac5e69"
      },
      "source": [
        "### Параметры по умолчанию - Свой классификатор\r\n",
        "knn = MyKNeighborsClassifier(n_neighbours=6, weights=\"uniform\", metric=\"euclidean\")\r\n",
        "knn.fit(X_train, y_train)\r\n",
        "y_pred = knn.predict(X_test)\r\n",
        "accuracy_score(y_test, y_pred)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "0.87"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 176
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "d2LO0062X5Vj",
        "outputId": "58ceed1a-946b-43a1-f5f6-14bcf061e29f"
      },
      "source": [
        "### Параметры по умолчанию - классификатор Sklearn\r\n",
        "knn = KNeighborsClassifier(n_neighbors=6, weights='uniform', metric='euclidean')\r\n",
        "knn.fit(X_train, y_train)\r\n",
        "y_pred = knn.predict(X_test)\r\n",
        "accuracy_score(y_test, y_pred)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "0.87"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 177
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "T-b1CruXMW3c",
        "outputId": "71af76d1-67a7-40c9-dc0a-5ace39122b55"
      },
      "source": [
        "### Манхэттенское расстояние - Свой классификатор\r\n",
        "knn = MyKNeighborsClassifier(n_neighbours=6, weights=\"uniform\", metric=\"manhattan\")\r\n",
        "knn.fit(X_train, y_train)\r\n",
        "y_pred = knn.predict(X_test)\r\n",
        "accuracy_score(y_test, y_pred)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "0.86"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 178
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "if48Wb2rMjB8",
        "outputId": "5254967d-4d11-4818-90ef-9ae9c0179c66"
      },
      "source": [
        "### Манхэттенское расстояние - классификатор Sklearn\r\n",
        "knn = KNeighborsClassifier(n_neighbors=6, weights='uniform', metric='manhattan')\r\n",
        "knn.fit(X_train, y_train)\r\n",
        "y_pred = knn.predict(X_test)\r\n",
        "accuracy_score(y_test, y_pred)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "0.86"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 179
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "x3NAsGITv6il",
        "outputId": "fa48025f-1e5d-4866-c32f-78d6bbf2bd92"
      },
      "source": [
        "### Веса, зависящие от расстояния - Свой классификатор\r\n",
        "knn = MyKNeighborsClassifier(n_neighbours=6, weights=\"distance\", metric=\"euclidean\")\r\n",
        "knn.fit(X_train, y_train)\r\n",
        "y_pred = knn.predict(X_test)\r\n",
        "accuracy_score(y_test, y_pred)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "0.88"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 180
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "35yuM4f5wUXM",
        "outputId": "7e18570f-d656-44d6-ab84-bbe0714e79b5"
      },
      "source": [
        "### Веса, зависящие от расстояния - классификатор Sklearn\r\n",
        "knn = KNeighborsClassifier(n_neighbors=6, weights='distance', metric='euclidean')\r\n",
        "knn.fit(X_train, y_train)\r\n",
        "y_pred = knn.predict(X_test)\r\n",
        "accuracy_score(y_test, y_pred)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "0.88"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 181
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "k6BQFJW0w6yl",
        "outputId": "ae325126-ead9-4a24-8642-783c25b92bdb"
      },
      "source": [
        "### Веса, зависящие от расстояния + Манхэттенское расстояние - Свой классификатор\r\n",
        "knn = MyKNeighborsClassifier(n_neighbours=6, weights=\"distance\", metric=\"manhattan\")\r\n",
        "knn.fit(X_train, y_train)\r\n",
        "y_pred = knn.predict(X_test)\r\n",
        "accuracy_score(y_test, y_pred)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "0.87"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 182
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "6fPyc8MkxBil",
        "outputId": "7dcc71f6-90c3-43c4-c10d-6dcf5908e53f"
      },
      "source": [
        "### Веса, зависящие от расстояния + Манхэттенское расстояние - классификатор Sklearn\r\n",
        "knn = KNeighborsClassifier(n_neighbors=6, weights='distance', metric='manhattan')\r\n",
        "knn.fit(X_train, y_train)\r\n",
        "y_pred = knn.predict(X_test)\r\n",
        "accuracy_score(y_test, y_pred)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "0.87"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 183
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "WN2pxwQy6Emo"
      },
      "source": [
        "2. Переберите параметр k от 1 до 10 для каждой модели из пункта выше (получится 4 графика по две линии на каждом)"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 295
        },
        "id": "ILiNTp8P6O_5",
        "outputId": "70959486-fc67-40cc-f0e8-5d637117f948"
      },
      "source": [
        "Accuracy_list = []\r\n",
        "for k in range(1, 11):\r\n",
        "  knn = MyKNeighborsClassifier(n_neighbours=k, weights=\"uniform\", metric=\"euclidean\")\r\n",
        "  knn.fit(X_train, y_train)\r\n",
        "  y_pred = knn.predict(X_test)\r\n",
        "  Accuracy_list.append(MyAccuracyScore(y_test, y_pred))\r\n",
        "\r\n",
        "\r\n",
        "Accuracy_list_sklearn = []\r\n",
        "for k in range(1, 11):\r\n",
        "  knn_sklearn = KNeighborsClassifier(n_neighbors=k, weights='uniform', metric='euclidean')\r\n",
        "  knn_sklearn.fit(X_train, y_train)\r\n",
        "  y_pred_sklearn = knn_sklearn.predict(X_test)\r\n",
        "  Accuracy_list_sklearn.append(accuracy_score(y_test, y_pred_sklearn))\r\n",
        "\r\n",
        "fig, ax = plt.subplots()\r\n",
        "ax.plot(range(1, 11), Accuracy_list, '-', label='My kNN', linewidth=2)\r\n",
        "ax.plot(range(1, 11), Accuracy_list_sklearn, '--', label='Sklearns kNN', linewidth=3)\r\n",
        "ax.legend()\r\n",
        "\r\n",
        "plt.title('Параметры по умолчанию')\r\n",
        "plt.grid()\r\n",
        "plt.xlabel('k')\r\n",
        "plt.ylabel('Accuracy, %')\r\n",
        "plt.show()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "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": "jHz7MYD2cHOx",
        "outputId": "72e72c4c-c9aa-4474-f21d-ef18b2f79aab"
      },
      "source": [
        "max(Accuracy_list_sklearn)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "0.91"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 185
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 295
        },
        "id": "PUTwzU4vRzlP",
        "outputId": "5ec17cbe-59af-4407-a666-6a033ce2cf9b"
      },
      "source": [
        "Accuracy_list = []\r\n",
        "for k in range(1, 11):\r\n",
        "  knn = MyKNeighborsClassifier(n_neighbours=k, weights=\"uniform\", metric=\"manhattan\")\r\n",
        "  knn.fit(X_train, y_train)\r\n",
        "  y_pred = knn.predict(X_test)\r\n",
        "  Accuracy_list.append(MyAccuracyScore(y_test, y_pred))\r\n",
        "\r\n",
        "Accuracy_list_sklearn = []\r\n",
        "for k in range(1, 11):\r\n",
        "  knn_sklearn = KNeighborsClassifier(n_neighbors=k, weights='uniform', metric='manhattan')\r\n",
        "  knn_sklearn.fit(X_train, y_train)\r\n",
        "  y_pred_sklearn = knn_sklearn.predict(X_test)\r\n",
        "  Accuracy_list_sklearn.append(accuracy_score(y_test, y_pred_sklearn))\r\n",
        "\r\n",
        "fig, ax = plt.subplots()\r\n",
        "ax.plot(range(1, 11), Accuracy_list, '-', label='My kNN', linewidth=2)\r\n",
        "ax.plot(range(1, 11), Accuracy_list_sklearn, '--', label='Sklearns kNN', linewidth=3)\r\n",
        "ax.legend()\r\n",
        "\r\n",
        "plt.title('Манхэттенское расстояние')\r\n",
        "plt.grid()\r\n",
        "plt.xlabel('k')\r\n",
        "plt.ylabel('Accuracy, %')\r\n",
        "plt.show()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "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/",
          "height": 295
        },
        "id": "wM2loZdj9mh3",
        "outputId": "8cf72af1-f2eb-4177-8092-78c32451fca0"
      },
      "source": [
        "Accuracy_list = []\r\n",
        "for k in range(1, 11):\r\n",
        "  knn = MyKNeighborsClassifier(n_neighbours=k, weights=\"distance\", metric=\"euclidean\")\r\n",
        "  knn.fit(X_train, y_train)\r\n",
        "  y_pred = knn.predict(X_test)\r\n",
        "  Accuracy_list.append(MyAccuracyScore(y_test, y_pred))\r\n",
        "\r\n",
        "Accuracy_list_sklearn = []\r\n",
        "for k in range(1, 11):\r\n",
        "  knn_sklearn = KNeighborsClassifier(n_neighbors=k, weights='distance', metric='euclidean')\r\n",
        "  knn_sklearn.fit(X_train, y_train)\r\n",
        "  y_pred_sklearn = knn_sklearn.predict(X_test)\r\n",
        "  Accuracy_list_sklearn.append(accuracy_score(y_test, y_pred_sklearn))\r\n",
        "\r\n",
        "fig, ax = plt.subplots()\r\n",
        "ax.plot(range(1, 11), Accuracy_list, '-', label='My kNN', linewidth=2)\r\n",
        "ax.plot(range(1, 11), Accuracy_list_sklearn, '--', label='Sklearns kNN', linewidth=3)\r\n",
        "ax.legend()\r\n",
        "\r\n",
        "plt.title('Веса, зависящие от расстояния')\r\n",
        "plt.grid()\r\n",
        "plt.xlabel('k')\r\n",
        "plt.ylabel('Accuracy, %')\r\n",
        "plt.show()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYgAAAEWCAYAAAB8LwAVAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4yLjIsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+WH4yJAAAgAElEQVR4nO3deXgV1fnA8e+bm+WShH0JyCKogOwgCCio4AqooK1VcStu2EWtFmu1WqXUve7WDZQfriBVq6gouKXaioogOyKrErawQ/bt/f0xk+Qm3JBLyM3cm/t+niePM2fmzH3nXMybOTNzjqgqxhhjTGVxXgdgjDEmMlmCMMYYE5QlCGOMMUFZgjDGGBOUJQhjjDFBWYIwxhgTlCUIY4wxQVmCqKdEZIOI5IpIlojsFpEPRKS913EZY6KHJYj67VxVTQXaANuApzyOxxgTRSxBxABVzQPeBLqXlolIkog8LCI/i8g2EXlORBoEbB8jIotEZJ+IrBWREW75lSKyUkT2i8g6Ebku1DhE5CER2eHW/VpEegZs+5eIbBWRvSLyhYj0CNg2TUQK3KuhXSLygojEu9smisirAfs+IyIqIse46w1E5BER+ck99n9Lz1NErnPLs0QkW0Q04DgbROR0dznePWZHd72jux4fEN89AXVnV9reWEReFJEtIrJJRO4REV8VbZQkIo+LyGb353ERSXK37XFjzRORYnc5S0QuDXKcYSKSISJ/cdt8Q+B+InK2iHzvfr8bRWRipfpDReQr9zM3isi4g7WniCx2Y8kVkZKA2P7i1jtRROa7deaLyIkBn9VNRP7nxpLlnlvp51X+fit/F9W1fZXHNtWzBBEDRCQZuAj4OqD4AaAL0Bc4BmgL3OXuPxB4GfgT0AQ4Gdjg1ssEzgEaAVcCj4nIcSGG8iLQwT3mZ8DfArZ9CHQGWgELgdcq1X3IvRrqDpwNjAhynl2AkZWKHwb6AycCzYBbgRIRSQGeAX7tHrdPiOdwUCIyHOhdqXgaUITTzv2AM4FrqjjEHcBgnO+lDzAQuBNAVZu4sf4GmKeqqe5P5bYq1RpogfPd/hqYLCJd3W3ZwBU438XZwG9F5Dz3HI7E+T6eAlq6sSxy6wVtT1Xt48Y2EtgcENt9ItIM+AB4EmgOPAp8ICLN3WPeDawEmrnHmFfF+RxUFW1fK8eOVZYg6rd3RGQPsBc4A/gHgIgIMB64WVV3qep+4D7gYrfe1cBUVf1YVUtUdZOq/gCgqh+o6lp1/AeYC5wUSjCqukpVcwBxixYGbJuqqvtVNR+YCPQRkcZBDuNz6+8Msu0+4O+lKyISB1wF/ME9h2JV/cr9jDigBIgPJfZQuO36EG6idcvSgFHATaqaraqZwGOUt3VllwKTVDVTVbfjJNHLDyOsv6pqvvtdfQBcCKCq6aq61P1+lwDTgVPcOpcAn6jqdFUtVNWdqrqomvY8mLOB1ar6iqoWqep04Afg3IB9fBzG76NgbV9bx45l1mj123mq2gTwA9cD/xGR1jh/FSYDC9wuhD3AR245QHtgbbADishIcbqHdrn1RuH8lRoSEXkG56/XS4DP3TKfiDwgTlfWPsqvVgKPe4v7eRtx/gqcX+m4g4GuwEsBxS3ccz/gXNykeDXwsojkEJCsDsOFwA6cq6NSRwIJwJaAtn4e50opmCOAnwLWf3LLamK3qmYHO5aIDBKRz0Vku4jsxbkqKW3vqr7/KtuzGpXPqTSWtu7yHcBRQI7bPoMr7XthQNvtqOIzgrV9KMc2B2EJIga4f+m9DRQDQ3H+R8oFerjdFk1UtbF7CQ7OL+GjKx/H7Qt/C6ebIc1NPrMpvyIIJZbf4SSnfwD/dosvAcYApwONgY6lHxlQ9WH38xoCiTjdX4EeAm5X1eKAsh1AXrBzcb0DFALDgVC7yaqSgHP18udK5RuBfKBFQFs3UtUeBxzBsRknqZTq4JbVRFO3Ky3YsV4HZgHtVbUx8Bzl7R30+6f69qxK5XMqjWUTgKquBRYDz7vf8deV9p1Z2nYE/2OkqrYP5djmICxBxABxjAGaAitVtQSYgnP/oJW7T1sROcut8iJwpYicJiJx7rZjcX4xJwHbgSIRGYnTnx74WSoiw6qIo6fbTSHucXLdTQ1xfonuxEke9x3kdIoBpfxqB+BUnH7w9wN3dM9zKvCoiBzhXqmc4CY6cO7DzFLVbw7yeaG6HPjK7a4JjGELTjfcIyLSyG3Po0XklKBHcbp67hSRliLSAqfL5NUq9g3F30QkUUROwrl39C+3vCGwS1Xz3HtOlwTUeQ04XUQudG8KNxeRviG0Z1VmA11E5BL3eBfh3Et6H8qu/s4Dbq/hOQZt+1o6dkyzBFG/vSciWcA+4F6cG7LL3W1/BtYAX7vdOp/gdNGgqt/i3oDGuX/xH+BIt1vmRmAmsBvnl8qs0g8T5z2L/cDSKuK53623E7gAGOuWv4zT5bAJWEHwv/Judc9lK86/2wcDtrXBuVkazC1uPPOBXW69OBEZgtM3/pcq6gFMF5EMyru85rnrpTc6pwfs2xT4axXHuQInua7AOf833ZiDuQf4Dljixr3QLauJre7nbcb5pf+b0ntJwO+ASSKyHycJzSytpKo/43QdTsBps0WU38QP2p4HC0JVd+Ikpwk43/2twDmqukNEEnD+WPmDqu6r4XkGbftaOnZME5swyNQWEbkMp9sqJv5aE5FPVPV0r+MIxr2Ke1VV23kdi4letfYEhzGqejhdIdGoNm5sGxOxrIvJmBpS1aq6tYypF6yLyRhjTFB2BWGMMSaoenMPokWLFtqxY0evwzgs2dnZpKSkVL9jjLD2qMjao5y1RUWH0x4LFizYoaotg22rNwmiY8eOfPfdd16HcVjS09MZNmyY12FEDGuPiqw9yllbVHQ47SEild9yL2NdTMYYY4KyBGGMMSYoSxDGGGOCsgRhjDEmKEsQxhhjggprghCRESKySkTWiMhtQbZ3cMek/15ElojIKLc8QUReEpGl4kxvGRNj+xhjTCQJ22Ou4sy5+zTOTGYZwHwRmaWqKwJ2uxNnrPdnRaQ7zrDAHYFfAUmq2sudLnOFiExX1Q3hitccxNZlkBNsArcgmh8DjdtWLNu0EPL3h1Y/rQekhDz/UGzK2QVbqxowt5LEFGg3oGLZ/q2wfVVo9Rs0hTaVZvHc/RPs3hBa/dQ0aHVsxbIda2DfptDqN2kPzY6qWLZtOU12L4F1IUxD0vxoaFxpvMJD+ffYqjukVnpF4OdvoCgvtPpH9AV/pYkR138BoY5g0X4gJDQoXy8uhJ++qriPLzG0Y9VAON+DGAisUdV1ACIyA2dSmMAEoThzG4MzUczmgPIUcSYebwAU4AxZbcLs7YUZzJi/0fkGXLfsnsTA/K+qrhTgxYa/Y07K6Apl9+78A50LQ/uF9I8mf2W+fwgALaWAISeVkOCL7Z7QdduzePCjH1i/OZdnfphH7/wF3Ln7jpDq/hx/JLe0eL5C2Sk5c/n9vkdDqr848TjubVZxeo7zsmZwSda0kOqn+0/nmSa3VCi7ct8zjMyZVUWNiv6dciHTG15VoeyPu+9hcP5/nWmAqjG14W/5KGVMhbK/77yZroUrQ/r8R5rcwTf+ijPq/jPzClqVZIZU/85mj/JjYvcKZa9t/QUJFIZU//ctprE9vnXZemrJPqZmXlhhnz1xTXki9TnC8VpIOBNEW5yZqUplAIMq7TMRmCsiNwApODOKgTNe/hhgC84EMjer6q7KHyAi43HmViYtLY309PRaDL/uZWVleXoODXI28/a3O1madwS5+MvKdycUOLP6hmDDrhy+3V7xq8pOLAq5M3NNZhbflpTX3/bkXG7quoei5NYHqVU/xRXnEb/xv/x2/YnszXcz9u5dJMXtd2aXCEFuQTHfbqj4fXTwZTtzsIVgb27hAfUH+HJDrr8ju4Bv91SsPyo+L+TfPFv25vHtzor1dx3Cv8efgvx7zDqMf48ABUklIc+huGLLPhZW+tVVkqQh11+8aS8ZWv5lN2E/Af9rAlBYXMKa3UVh+d3h9ZvUY4FpqvqIiJwAvCIiPXGuPopx5rJtCnzpjr2/LrCyqk4GJgMMGDBAo/3NSq/fDi14/1Ze5XlKkoSfB/6Vbd3GAdB20QnszQztN8Llxw7mnA4Vp/1t8/Vg9u6ragrmiv7QZxBXpQ1kV3YBf3xjIY22z2fw3sfhjInED7nhkM4nquXtI2faL0jeOp9Li7ax+JirOalZNv369iVlRzJ7F54Q0mGaprTljSEVv4/Gm/LZu/yAydeCOqZZT94YULF+83Xb2Lvm55DqDzxiMG/0rFg/beUP7N0YWhfPiI6D6d+lYv22iwazaW0OqQ0bVlv/0mMHM6rSv8e235zA3r2hdWPe0Hsw41pX/LvW/8Vg9uaF1uU68fhB5DbtWqEs9+MTyNOikOo/edJgChuUd3H5CvaxN73id+9LbMQfmjYIy++OsI3m6v7Cn6iqZ7nrtwOo6v0B+ywHRqjqRnd9Hc6k4ncDX6vqK275VOAjVZ1JFQYMGKA21Mbh2Tf5bBpt/q+zctFr0O0cz2IBmP3aPzl99d0k4vzPVDj8bhJO+aOnMdWJ3D3k/N/5JGeWTzeRP+5j5m3IseElXF7/vxJpDnOojQWqOiDYtnB27s4HOotIJxFJBC4mYHpK18/AaW6Q3XAunra75ae65Sk4SeMHTFgl7Axo4lbdvAvE1TDtKIpa9ytbT/j8bxR89oCHEdWBnF3kvHhOheRQdMY9JHUc6GFQJlaFLUGoahFwPTAHWInztNJyEZkkIqV3MScA14rIYpz5fcepc0nzNJDqXmHMB/4v2ITkphbl7KJB/g4AiuKSoGlHb+MBiuOTSb7qXXKOOLGsLPGL+ymY+7fQnwKJJtk7yJkyiuQd5U8oFY14KLa61kxECes9CFWdjfPoamDZXQHLK4AhQepl4TzqaupKZvlTHTmNj6FRXIh3AcMtMYXkcW+R88pFJG/8win66lHyiwtIGnEPSIh3+yJdViY5L5xN8p4fAShB0LMfI/74Kz0OzMSy2H5+0JTLLH/62Ne6+0F29EBiMslX/IucjqeXFSV980/y3/9z/biS2LeFnMkjApJDHIx5Gp8lB+MxSxAGgLxNy8qWk9v29DCSKiT4Sb5sOjlHjSgrSlrwPHnv3AQlJR4Gdpj2ZpAz+SyS960FoBgfnP88cf0u9TgwYyxBGFfBluVly9Iqwq4gSsUnknzpq+R2PresyL94Gnt/+MzDoA7P/EWLkf1bACc5yAVTietzYTW1jKkbliAMqJK0+8fy9Qh4gqlKvgQaXDyNvGN/CcDfCi/nlx8lkLkvxKEPIsjspVsYO0e4pnACeXEpxF30MnE9z/M6LGPKWIIwkLWNpMK9ABT4Ug4cuybS+OLxXziFvb+YzlctLmRNZhYXTf6aLXtzvY4sZO8u2sQN07+nqETpedIYkm5Zjnj83okxlVmCMFCQzaLE/mzWZuQ17RIdTwbF+WjcexTTxw+me5tGrN+RzUXPf82mbduhqMDr6KqWuZIPv5jHzW8sorhEufHUY7htxLFIclOvIzPmAJYgDNrsKK4o+DMn5v+TvEve9TqcQ9IsJZHXrx1E73aNydy1m23PjSHn1bFQGIFdTluXkvfCSHp9ehltdDsTzujCH8/sikRDQjYxyRKEYdu+fPblFdEkOYGWTRtVXyHCNElO5NUr+/FGwyc4TpeTvOETcl6+EApyvA6t3ObvyX9hFP6C3bSTHbzb4hluGH6011EZc1CWIAw/bnMGTuuS1jBq/5ptlJxMt+NPLVtP3vgfcqb9EgqyPYzKtXE++VPPIanIGbE+Pz6VFhc/A3H2v5+JbPYv1JQliK5p1Y+OGbFESDzzbgpOLp98MHnzV+RMPS/0yWHC4ad5FEwbTVJRFgD5CY1Juur9AyfxMSYCWYKIdfs2033h37jM9zEn+td7Hc1hSzz1NgqHl43mQvLWb8l5cTTk7qn7YNZ/SeFL55FY7HR15SU2JenqD+CIftVUNCYyeD0fhPHaliWcuPsdTkyAfetXAudWWyXSJZwygUJfIgmf3AlAcuZCZ4TUq2ZBcrM6iUHXfEbR6xeTUJIPQF5Sc/xXfxDZ75gYU4ldQcS4km3lYzAlHtHDw0hqV8LQGyga8VDZevKOpeRMGQXZIc6tfRj0xzkUv3ZRWXLI9bfCf+0cSw4m6liCiHG5m8qHlvZH4hhMhyF+8HUUnf04Je78jsm7V7LjXzeG9TNVlWUfTiFenXcxchu0ocG1H0GLzmH9XGPCwRJEjAu8gqBl/fsLN/74K9HRT1OC8GNJW0avPZf/rt4Rls8qKVHufGcZv9hyOZ+W9CcnpR0Nxs+B5vY4q4lOdg8ilhUXkbx3bfl6q2O9iyWMfMddSrG/ETMWp7B5cS5XvTSf5y/vz/Cuoc2THYriEuX2t5cw87sMEuMTib/4JZLbJ0DD1rX2GcbUNbuCiGW71+PTQgBy/K2gQf0d7sHX/VzuvGg4lw3uQEFRCde9vICPV2yDovzDPnbRxgXcMnMRM7/LwJ8Qx/+NO55Ture35GCiniWIWBYwSVBR8/p59RAoLk74+5ieXDmkIwXFJbzz2jNkP9Yfdq2r8TGL5k8l/sVT6bTscZITfUy7ciBDjmlRi1Eb4x1LEDGseGt5gvC37eVhJHVHRLjrnO78o1cGj8c/RUr2RnInnwU7Vh/ysYrmPUf8BzcDcGP8O3w4dC2Dj2pe2yEb4xlLEDEs8AmmxDYROklQGIgIFwzqDD7nFlyDvEzyppxVYV7u6hR++QTxc/5ctp7dog9HnnRJrcdqjJcsQcQwzfyhfCXGntGXY04l4fI3KYzzA+DP30neCyNh69JqakJh+sMkfFr+tnZOq/6kXPNevb6HY2KTJYgYln7EeJ4oOp9VTYdBi65eh1P3Op1Mwq/fcSZJAvwFu8l/YRRs/j74/qoUfHofCel/LyvKaTOI5KvfBX/juojYmDplCSKGzS4+nseKfsWKk5+GpFSvw/HGkSeQeOUs8uOd808q2kf+1HNg4/yK+6lSMHciiV8+WFaU03YIyVf+G5KieJBDYw7CEkQMCxzmO6a1G0DSVe+Tn+BcBSQVZVEwbTT8NM/Zrkr+7NtJnPd4WZWcDsNIHvcWJKZ4EbExdSKsCUJERojIKhFZIyK3BdneQUQ+F5HvRWSJiIxyyy8VkUUBPyUi0jecscaavMJiNuzMIU7g6JYxevUQ6Ih+JF39AXmJzn2ExOIcCl7+BWRtZ9+W1RR/N61s15yOp5N8+RuQ0MCjYI2pG2FLECLiA54GRgLdgbEiUvlRmTuBmaraD7gYeAZAVV9T1b6q2he4HFivqovCFWssWpeZRXGJ0rFFCv4En9fhRIbWvfBf8yF5Sc6jqhPzxnJP+nYu/tc2fp33J3JJIufoUSRfNh0S/B4Ha0z4hXOojYHAGlVdByAiM4AxQMDgPyhQOsdlY2BzkOOMBWaEMc6Y1HbGacxKLCKbLpDTu86GwY54rbrhv3YO3/93Nm9825ni/zpzZHRq0Z+s82bTslMf8CV4HKQxdUNUNTwHFrkAGKGq17jrlwODVPX6gH3aAHOBpkAKcLqqLqh0nLXAGFVdFuQzxgPjAdLS0vrPmBHdeSQrK4vU1PB398QV53HylxcBUEwc/z15JhoXeb/06qo9qjJ/axHPL86nVYpw6wA/Tfze3rLzuj0iibVFRYfTHsOHD1+gqkGnOPR6sL6xwDRVfURETgBeEZGeqloCICKDgJxgyQFAVScDkwEGDBigw4YNq6OwwyM9PZ06OYdNC+BLZzG3YUdOOfWM8H9mDdRZe1RhGHDNuQWk+uNJ8Hn/PIfX7RFJrC0qCld7hPNf/SagfcB6O7cs0NXATABVnQf4gcCBbC4GpocxxtgUwy/IHaqmKYkRkRyM8UI4/+XPBzqLSCcRScT5ZT+r0j4/A6cBiEg3nASx3V2PAy7E7j/UusIty8uWG8TIGEzGmEMXtgShqkXA9cAcYCXO00rLRWSSiIx2d5sAXCsii3GuFMZp+U2Rk4GNpTe5Te3J3VTeY+dLsysIY0xwYb0HoaqzgdmVyu4KWF4BDKmibjowOJzxxar4nYFdTLEzSJ8x5tBY52qsyd1Dct42AIolAZod5XFAxphIZQki1mwvv3rIaXx02ZDXxhhTmSWIWBMwi1xcmnUvGWOqZgkixuRvKb+CaNC2p4eRGGMinSWIGLOk560MzX+CSY3uJq7HeV6HY4yJYNYBHWNWbcsmQ1uyr30/aH601+EYYyKYXUHEmNI5ILrG+hwQxphqWYKIMau2upMEtbYEYYw5OEsQMUR3rEa3LiORQrqk2UiYxpiDs3sQMSTviyeYqa9Q5I/Dt/YJOO4Kr0MyxkQwu4KIIaWD9MVTgjRq63E0xphIZwkiVqji3/Vj+bqNwWSMqYYliFixbxOJxVkA5Mc3goatPQ7IGBPpLEHEisyVZYv5zbqAiIfBGGOigSWIGKHbysdgSmzTw8NIjDHRwhJEjMjdtLRs2W+zyBljQmAJIkYUbS3vYrJ5qI0xobAEEQtKimmwd035ektLEMaY6lmCiAW7N5BQkgdAblILSGnucUDGmGhgCSIWFOayOL43O7Uhhc27eh2NMSZK2FAbMaC4VQ8uzPsL+UUlLBk71OtwjDFRwq4gYsDPu3LILyrhiMZ+GjVs7HU4xpgoYQkiBpQO8d3Z5oAwxhwCSxAxoGySIJsDwhhzCMKaIERkhIisEpE1InJbkO0dRORzEfleRJaIyKiAbb1FZJ6ILBeRpSLiD2es9dbuDfRaej8X+T5noD/D62iMMVEkbDepRcQHPA2cAWQA80VklqquCNjtTmCmqj4rIt2B2UBHEYkHXgUuV9XFItIcKAxXrPXapgUM3/MWwxNg3/q1OF+HMcZUL5xXEAOBNaq6TlULgBnAmEr7KNDIXW4MbHaXzwSWqOpiAFXdqarFYYy13ireWp6PG7Tt6WEkxphoE87HXNsCGwPWM4BBlfaZCMwVkRuAFOB0t7wLoCIyB2gJzFDVhyp/gIiMB8YDpKWlkZ6eXpvx17msrKxaP4ejln1JB3d5zZ44tkVRG4WjPaKZtUc5a4uKwtUeXr8HMRaYpqqPiMgJwCsi0tONayhwPJADfCoiC1T108DKqjoZmAwwYMAAHTZsWJ0GX9vS09Op7XPImn9j2XK3U35BtzZ9avX44RSO9ohm1h7lrC0qCld7hLOLaRPQPmC9nVsW6GpgJoCqzgP8QAucq40vVHWHqubg3Js4Loyx1k+FuaRk/wxACXHQoovHARljokk4E8R8oLOIdBKRROBiYFalfX4GTgMQkW44CWI7MAfoJSLJ7g3rU4AVmEOzfRWCApCT0h4SGngckDEmmoSti0lVi0Tkepxf9j5gqqouF5FJwHeqOguYAEwRkZtxbliPU1UFdovIozhJRoHZqvpBuGKtt7b/ULaoNsS3MeYQhZwgRGQwzk1lP/C4qr5TXR1VnY3TPRRYdlfA8gpgSBV1X8V51NXUUOGW5SS4y8k2SZAx5hBVmSBEpLWqbg0o+iNwPiDAN0C1CcJ4K3fTsrIE4Wvd3dNYjDHR52BXEM+JyELgIVXNA/YAFwAlwL66CM4cHt8Om0XOGFNzVd6kVtXzgO+B90XkCuAmIAloDpxXN+GZGlPl0zbX8WzRuaxvdhI0O9rriIwxUeag9yBU9T0RmQ38Dvg3cK+qflEnkZnDI8LbxUNIL+pCp1P70yk+0euIjDFRpsorCBEZLSKfAx8By4CLgDEiMkNE7M/RKPCjO8x3l7RUjyMxxkSjg11B3IMznlIDYI6qDgQmiEhn4F6c9xpMhNqXV8jmvXkkxsdxZPMUr8MxxkShgyWIvcAvgGQgs7RQVVdjySHird6WBUDnVqn44sTjaIwx0ehgCeJ8nLGSCoFL6iYcU1s6zxzGW4kJFJR0hfy+kGSTBRljDk2VCUJVdwBP1WEsprZk76RR9gb6x0Hhvo2QkOx1RMaYKGRTjtZH28vff8htfDTE+TwMxhgTrSxB1EeZ5QnC3qA2xtSUJYh6KG/TsrJlG4PJGFNTh5wgROQTEflQRM4JR0Dm8BVsWV62LGl2BWGMqZmaDPd9BdAGGFzLsZjaoErSrlXl6zYGkzGmhqpNECJyLvCBqpYAqOpmYDOwIMyxmZrYv5WkImcsxQJfKomN2nockDEmWoXSxXQRsFpEHhKRY8MdkDlMmeUT7+U17QxiL8kZY2qm2gShqpcB/YC1wDQRmSci40XE3ryKQBrwBFNCmx4eRmKMiXYh3aRW1X3Am8AMnPsP5wMLReSGMMZmaiB3c/kVhL9tTw8jMcZEu2oThDuq67+BdCABGKiqI4E+OHNKmwgyv+ddDM9/hEea3Il0GeF1OMaYKBbKU0y/BB6rPA+EquaIyNXhCcvU1I+ZOazXNuzpOBiadfI6HGNMFAslQUwEtpSuiEgDIE1VN6jqp+EKzNTMj9vcOSBa2y0iY8zhCeUexL9w5qEuVeyWmQhUmiC6plmCMMYcnlASRLyqFpSuuMs2f2UEKsn8kYJtP+Kj2GaRM8YctlASxHYRGV26IiJjgB3hC8nUVO4n9/Gh72ZW+K+iybr3vQ7HGBPlQkkQvwH+IiI/i8hG4M/AdaEcXERGiMgqEVkjIrcF2d5BRD4Xke9FZImIjHLLO4pIrogscn+eO5STilUl7jsQSRRCoyM8jsYYE+2qvUmtqmuBwSKS6q5nhXJgEfEBTwNnABnAfBGZpaorAna7E5ipqs+KSHdgNtDR3bZWVfuGfCaxrriI5L1ry9db2kvvxpjDE9JgfSJyNtAD8Is7dIOqTqqm2kBgjaquc48xAxgDBCYIBRq5y41xxngyNbFrHT4tBCDHn0ZygyYeB2SMiXahDNb3HJAMDAdeAC4Avg3h2G2BjQHrGcCgSvtMBOa6b2SnAKcHbOskIt8D+4A7VfXLILGNB8YDpKWlkZ6eHkJYkSsrK6vG59Ay83+UDqyxK6EN30Z5W8DhtUd9ZO1RztqionC1RyhXECeqam8RWaKqfxORR4APa+nzxwLTVPURETkBeEVEeuK8d9FBVXeKSH/gHRHp4Q75UUZVJwOTAQYMGKDDhtqoVVsAABsWSURBVA2rpbC8kZ6eTk3PofizeWXLad2G0C7K2wIOrz3qI2uPctYWFYWrPUK5SZ3n/jdHRI4ACnHGY6rOJqB9wHo7tyzQ1cBMAFWdB/iBFqqar6o73fIFOAMFdgnhM2NW7qalZcs2SJ8xpjaEkiDeE5EmwD+AhcAG4PUQ6s0HOotIJxFJBC4GZlXa52fgNAAR6YaTILaLSEv3JjcichTQGVgXwmfGroBRXG2SIGNMbThoF5OIxAGfquoe4C0ReR/wq+re6g6sqkUicj0wB/ABU1V1uYhMAr5T1Vk4g/1NEZGbcW5Yj1NVFZGTgUkiUojzFvdvVHXX4ZxovVaYR/L+n8rXW3b1LhZjTL1x0AShqiUi8jTOfBCoaj6QH+rBVXU2zqOrgWV3BSyvAIYEqfcW8FaonxPzdq4mjmIAspPbk5KY4nFAxpj6IJQupk9F5JciNjVZxCoqYKmvG/s0mWJ7/8EYU0tCeYrpOuCPQJGI5AECqKo2Ong1U1fy0voyJuevgLLiwpO8DscYU0+E8ia1DQsa4dZtz6ZE4aiWqfhTGnsdjjGmngjlRbmTg5VXnkDIeMeG+DbGhEMoXUx/Clj24wyhsQA4NSwRmUO2qnSSIEsQxphaFEoX07mB6yLSHng8bBGZQ7NjNb1XPMr5cc04rkEj7H1CY0xtCWmwvkoyAHsTK1Js/IaR+95gZCLsX78RhtpNamNM7QjlHsRTOC+xgfNYbF+cN6pNBCjcspwEdzm5bU9PYzHG1C+hXEF8F7BcBExX1f+FKR5ziHI3lScIX+vunsZijKlfQkkQbwJ5qloMzkRAIpKsqjnhDc2EIn7nD+UrrSxBGGNqT0hvUgMNAtYbAJ+EJxxzSHJ3k5y3DYAiSYSmnTwOyBhTn4SSIPyB04y6y8nhC8mELLP86iG38VHgq8kzB8YYE1woCSJbRI4rXXEn8MkNX0gmZJnls7fGpVn3kjGmdoXyJ+dNwL9EZDPOOEytgYvCGpUJSf6W5SS5yw3a9vI0FmNM/RPKi3LzReRYoHSSgVWqWhjesEwo8jeXJwi7gjDG1LZqu5hE5PdAiqouU9VlQKqI/C78oZmDUiWxwhNM9u6iMaZ2hXIP4lp3RjkAVHU3cG34QjIh0RI+bP0bXigayc/NT4LG7auvY4wxhyCUexA+ERFVVXDegwASwxuWqVacjxlFw/imqDedzxpIh7hQcr0xxoQulATxEfCGiDzvrl/nlhkPqaoN822MCatQEsSfgfHAb931j4EpYYvIhGR7Vj67cwpp5I8nrVFS9RWMMeYQVdsvoaolqvqcql6gqhcAK4Cnwh+aOZgftzrvLnZJa4hNF26MCYeQXr0VkX7AWOBCYD3wdjiDMtVQpffbw5ie0BCKjoXC4yDB73VUxph6psoEISJdcJLCWGAH8AYgqjq8jmIzVdm3iUa5GznBB/l7MyDeupiMMbXvYFcQPwBfAueo6hoAEbm5TqIyB5e5smwxv1lXkqyLyRgTBge7B/ELYAvwuYhMEZHTcIbaCJmIjBCRVSKyRkRuC7K9g4h8LiLfi8gSERkVZHuWiNxyKJ9b3+m28jGYEtv08DASY0x9VmWCUNV3VPVi4Fjgc5wxmVqJyLMicmZ1B3bfl3gaGAl0B8aKSOXxIO4EZqpqP+Bi4JlK2x8FPgz1ZGJFzqalZct+m0XOGBMmoTzFlK2qr6vquUA74HucR1+rMxBYo6rrVLUAmAGMqXx4oJG73BjYXLpBRM7DuSG+PITPiinFW8uvIGySIGNMuIj7gnTtH1jkAmCEql7jrl8ODFLV6wP2aQPMBZoCKcDpqrpARFJx3rc4A7gFyFLVh4N8xnicdzRIS0vrP2PGjLCcS13JysoiNTX14DtpMUO+uIgEd7zE/w55haKERgevE6VCao8YYu1RztqiosNpj+HDhy9Q1QHBtnk9w8xYYJqqPiIiJwCviEhPYCLwmKpmHewZf1WdDEwGGDBggA4bNiz8EYdReno61Z7DzrXwHyc55Ca1YOgZo8MfmEdCao8YYu1RztqionC1RzgTxCYgcAS5dm5ZoKuBEQCqOk9E/EALYBBwgYg8BDQBSkQkT1X/GcZ4o0PAE0yFzY+tMBesMcbUpnAmiPlAZxHphJMYLgYuqbTPz8BpwDQR6Qb4ge2qelLpDiIyEaeLyZIDULJtRdmNo6Qj7AkmY0z4hG0IUFUtAq4H5gArcZ5WWi4ik0SktF9kAnCtiCwGpgPjNFw3ReqJnE3LypaTjrAnmIwx4RPWexCqOhuYXansroDlFcCQao4xMSzBRan/9byHR5edwHlt9/Hbo4Z5HY4xph7z+ia1OUSrduSzSjuw56ijoEkHr8MxxtRjNstMlFnlzgHRxeaAMMaEmSWIKPPjVksQxpi6YQkiihRsW0XhjvXESQnHtLKXhIwx4WX3IKJI/od/JT1xDrn4abDhZehyltchGWPqMbuCiCKy3XlJrgF50LC1x9EYY+o7SxDRojCXlOyNAJQQBy26eByQMaa+swQRLbavQnDeIcxJ7QAJNsiGMSa8LEFEi4AxmLRlNw8DMcbECksQUaJwa/m0GMk2SZAxpg5YgogSeRnlYzD5WtskQcaY8LMEESXidv5QvmKzyBlj6oAliGiQt4+U3C0AFEs8ND/a44CMMbHAEkQ02F5+9ZDT6CjwJXgYjDEmVliCiAbFBayUo8nRJKSVPcFkjKkbNtRGFNjXehAjc/+OPx6W/2Ko1+EYY2KEXUFEgdXuEN/HpDXC16Cxx9EYY2KFJYgosGprFgBdWtkQ38aYumMJIgr8WDpJUGtLEMaYumP3ICLdthX0Wf1Pzo1rRt/kVoA94mqMqRuWICLdT//j/P2vc34iZK/LhOOP9zoiY0yMsC6mCJe7aWnZcnK7Xh5GYoyJNZYgIlzhlvJB+uwdCGNMXbIEEclUSdq1qnzdEoQxpg6FNUGIyAgRWSUia0TktiDbO4jI5yLyvYgsEZFRbvlAEVnk/iwWkfPDGWfE2r+VpCLnCaYCXyo0autxQMaYWBK2m9Qi4gOeBs4AMoD5IjJLVVcE7HYnMFNVnxWR7sBsoCOwDBigqkUi0gZYLCLvqWpRuOKNSJnlTZXXrAuJIh4GY4yJNeG8ghgIrFHVdapaAMwAxlTaR4FG7nJjYDOAquYEJAO/u1/M0YAEkdCmh4eRGGNiUTgfc20LbAxYzwAGVdpnIjBXRG4AUoDTSzeIyCBgKnAkcHmwqwcRGQ+MB0hLSyM9Pb0Ww697WVlZFc6h4/LP6Ogub8xOYnOUn9+hqtwesc7ao5y1RUXhag+v34MYC0xT1UdE5ATgFRHpqaolqvoN0ENEugEviciHqpoXWFlVJwOTAQYMGKDDhg2r6/hrVXp6OoHnsG/pXWXLXYaMpstRp3gQlXcqt0ess/YoZ21RUbjaI5xdTJuA9gHr7dyyQFcDMwFUdR5Od1KLwB1UdSWQBcTWRMwlJTTY82P5us0iZ4ypY+FMEPOBziLSSUQSgYuBWZX2+Rk4DcC9UvAD29068W75kcCxwIYwxhp5SoqY1eq3vFR0BhktT4HUll5HZIyJMWHrYnKfQLoemAP4gKmqulxEJgHfqeosYAIwRURuxrkRPU5VVUSGAreJSCFQAvxOVXeEK9aIFJ/IS4WnsaRoAN1GnUA7r+MxxsScsN6DUNXZOI+uBpbdFbC8AhgSpN4rwCvhjC3SlZQoq7e5w3ynpXocjTGHprCwkIyMDPLy8qrfuQYaN27MypUrw3LsaBRKe/j9ftq1a0dCQuhTFnt9k9pUIWN3LrmFxaQ1SqJJcqLX4RhzSDIyMmjYsCEdO3ZEwvD+zv79+2nY0Ia/L1Vde6gqO3fuJCMjg06dOoV8XBtqI0KtKp0DIs3+JzDRJy8vj+bNm4clOZhDJyI0b978kK/o7AoiEhUXMnDWcF5OaEZiQU8oOR7iLJeb6GLJIbLU5PuwBBGJdq2jcd4mTvZtImdvpiUHY4wn7DdPJMosv9lU1OJYDwMxJnqJCJdddlnZelFRES1btuScc84J+RgTJ07k4YcfPqB83LhxtG3blvz8fAB27NhBx44dAdiwYQMiwlNPPVW2//XXX8+0adNqdiIesgQRgYq3lY/B5G9rkwQZUxMpKSksW7aM3NxcAD7++GPatq29EZF9Ph9Tp04Nuq1Vq1Y88cQTFBQU1NrnecG6mCJQbsZSSh9sTbRB+kyU63jbB2E57oYHzq52n1GjRvHBBx9wwQUXMH36dMaOHcuXX35JSUkJXbt25auvvqJly5aUlJTQpUsX5s2bR8uWwV9KnTJlCm+//TZvv/02ADfddBOPPfYY11577QH7tmzZkiFDhvDSSy8F3R4t7AoiEgV0MdHSupiMqamLL76YGTNmkJeXx5IlSxg0yBkvNC4ujssuu4zXXnsNgE8++YQ+ffpUmRz++c9/8v777/POO+/QoEEDADp06MDQoUN55ZXgr2z9+c9/5uGHH6a4uDgMZ1Y37Aoi0hTmkZz1MwCKIC27ehyQMYcnlL/0D9X+/ftD2q93795s2LCB6dOnM2rUqArbrrrqKsaMGcNNN93E1KlTufLKK4Me4+WXX6Z9+/a88847B7xkdvvttzNmzBjOPvvAczzqqKMYNGgQr7/+eohnFXnsCiLS7FxNHM5fHNkp7SAxxeOAjIluo0eP5pZbbmHs2LEVytu3b09aWhqfffYZ3377LSNHjgxav1evXmzYsIGMjIwDtnXu3Jm+ffsyc+bMoHX/8pe/8OCDD6IanVPaWIKINAHdSyUtbA5qYw7XVVddxd13302vXgc+8HHNNddw2WWX8atf/Qqfzxe0fr9+/Xj++ecZPXo0mzdvPmD7HXfcEfRJJ4Bjjz2W7t2789577x3eSXjEEkSEKdq6vGw5uW1sjXBuTDi0a9eOG2+8Mei20aNHk5WVVWX3UqmhQ4fy8MMPc/bZZ7NjR8VxQ3v06MFxxx1XZd077rgj6NVHNLB7EBEmN2MppYNrxNsTTMbUWFZW1gFlw4YNqzCxzuLFi+nTpw/HHhv8YZCJEyeWLZ911lmcddZZAAe801D6ZBNAx44dWbZsWdl6nz59KCkpqcEZeM8SRIT5vNeDPL96Dr9qv59xR57odTjG1FsPPPAAzz77bNmTTOZA1sUUYVbsKGa5dmJ3519CoyO8DseYeuu2227jp59+YujQoV6HErEsQUSY1e4orl1b2yiuxhhvWYKIMDbMtzEmUtg9iAgi+zMo3K0k+prTsXmy1+EYY2KcJYgIcvSPU/jGv4j9kkL8z29Ap5O8DskYE8OsiymCNMpxhthoqNnQsI3H0RgT3e6991569OhB79696du3L9988w3gPIZa+V0GgNRU7+d+T09PDzoc+bRp04iLi2PJkiVlZT179mTDhg1ly7/85S/Ltr355puMGzfusOOxBBEpcnfTqHgXAEWSCM1CnzfWGFPRvHnzeP/991m4cCFLlizhk08+oX379nXy2eEanK9du3bce++9VW5fsGABK1asqHJ7TViCiBSZP5Qt5jY+GuKCv/ZvTFT6/H6Y2Di0n1lB3nqedWOFfRK/euSgH7dlyxZatGhBUlISAC1atOCIIyo+Np6bm8vIkSOZMmXKAfX/8Y9/cPzxx9O7d2/uvvvusvLzzjuP/v3706NHDyZPnlxWnpqayoQJE+jTpw/z5s0jNTWVO+64gz59+jB48GC2bdsGwL/+9S969uxJnz59OPnkkw96DvPnz6dfv36sXbsWgHPOOYfly5ezatWqoPtPmDDhoAmkJixBRIrM8swfl9bdw0CMiX5nnnkmGzdupEuXLvzud7/jP//5T4XtWVlZnHvuuYwdO/aA+Rrmzp3L6tWr+fbbb1m0aBELFizgiy++AGDq1KksWLCA7777jieffJKdO3cCkJ2dzaBBg1i8eDFDhw4lOzubwYMHs3jxYk4++eSyJDRp0iTmzJnD4sWLmTVrVpXxf/XVV/zmN7/h3Xff5eijjwacIcpvvfVW7rvvvqB1LrzwQhYuXMiaNWtq1mhBWIKIEHmby1/Nb9DOZpEz5nCkpqayYMECJk+eTMuWLbnooosqDI8xZswYrrzySq644ooD6s6dO5e5c+fSr18/jjvuOH744QdWr14NwJNPPll2VbBx48aycp/PV+EeQGJiYtm9hP79+5fdKxgyZAjjxo1jypQpVXZFrVy5kvHjx/Pee+/RoUOHCtsuueQSvv76a9avX39APZ/Px5/+9Cfuv//+0BuqGmF9iklERgBPAD7gBVV9oNL2DsBLQBN3n9tUdbaInAE8ACQCBcCfVPWzcMbqtYLNK/C7y3YFYeqd4bc7PzU1+knnx1Wwfz9J1VTx+XxlYy/16tWLl156qezG7ZAhQ/joo4+45JJLEJEK9VSV22+/neuuu65CeXp6Op988gnz5s0jOTmZYcOGkZeXB4Df768wGmxCQkLZcX0+H0VFRQA899xzfPPNN3zwwQf079+fBQsW0Lx58wqf06ZNG/Ly8vj+++8P6BaLj49nwoQJPPjgg0HP+fLLL+f++++nZ8/aGegzbAlCRHzA08AZQAYwX0RmqWrgXZQ7gZmq+qyIdAdmAx2BHcC5qrpZRHoCc4Dam0w2QG5BMYU/fkLimg9D2r8orS/5vSqOK5+46j0SNv43pPqFR55MQeeKk4skLX0d/46l5QU2i5wxh2XVqlXExcXRuXNnABYtWsSRRx5Ztn3SpElMmjSJ3//+9zzzzDMV6p511ln89a9/5dJLLyU1NZVNmzaRkJDA3r17adq0KcnJyfzwww98/fXXhxzX2rVrGTRoEIMGDeLDDz9k48aNBySIJk2a8OKLL3LGGWeQkpJSYXBBgHHjxvHQQw8FnTQpISGBm2++mQceeIBTTz31kOOrLJxXEAOBNaq6DkBEZgBjgMAEoUAjd7kxsBlAVb8P2Gc50EBEklQ1v7aDfHneBnbO/Td/SZge0v7vFJ/ITe9WnJbwr/FvcnV8aAnmpe8yub+o4t8+TyXM4FxfNgAFvmQSG9fN0xbG1FdZWVnccMMN7Nmzh/j4eI455pgKN5UBnnjiCa666ipuvfVWHnroobLyM888k5UrV3LCCScATnfVq6++yogRI3juuefo1q0bXbt2ZfDgwYcc15/+9CdWr16NqnLaaafRp0+foPulpaXx/vvvM3LkSKZOnVphW2JiIjfeeCN/+MMfgta9+uqrueeeew45tmAkXDMdicgFwAhVvcZdvxwYpKrXB+zTBpgLNAVSgNNVdUGQ4/xGVU8P8hnjgfEAaWlp/WfMmHHIcX7yUyFN177FhLjQEsQHJSdyW/kpAHCrvMLlcaEliP8rOZtH9dIKZQ/Jk4yMc/4a+an5yazvNSGkY9V3WVlZEfFseqSIpvZo3LgxxxxzTNiOX1xcXOUEP7Eo1PZYs2YNe/furVA2fPjwBao6INj+Xr9JPRaYpqqPiMgJwCsi0lNVSwBEpAfwIHBmsMqqOhmYDDBgwACtfCkWimEAW1rBxv4h7X9286M5++hKl24/NYNtp4VU/8rWvbiyQ6W/PNYkwa71rFi/he7nT+BIm2YUcPp8a/Kd1lfR1B4rV66kYcPwjSe2f//+sB4/2oTaHn6/n379+oV83HAmiE1AYF9JO7cs0NXACABVnScifqAFkCki7YB/A1eo6towxglt+jg/NXXkCc5PTR3jXBxl5qTT3ZKDMSZChPMx1/lAZxHpJCKJwMVA5Qd/fwZOAxCRboAf2C4iTYAPcJ5q+l8YYzTGhEm4uq9NzdTk+whbglDVIuB6nCeQVuI8rbRcRCaJyGh3twnAtSKyGJgOjFPnLK4HjgHuEpFF7k+rcMVqjKldfr+fnTt3WpKIEKrKzp078fv91e8cIKz3IFR1Ns6jq4FldwUsrwCGBKl3D1A7t+GNMXWuXbt2ZGRksH379rAcPy8v75B/2dVnobSH3++nXbt2h3Rcr29SG2PqoYSEBDp1Ct+Ak+np6Yd0s7W+C1d72FAbxhhjgrIEYYwxJihLEMYYY4IK25vUdU1EtgM/eR3HYWqBMw6VcVh7VGTtUc7aoqLDaY8jVbVlsA31JkHUByLyXVWvvMcia4+KrD3KWVtUFK72sC4mY4wxQVmCMMYYE5QliMgyufpdYoq1R0XWHuWsLSoKS3vYPQhjjDFB2RWEMcaYoCxBGGOMCcoSRAQQkfYi8rmIrBCR5SISfC7BGCIiPhH5XkTe9zoWr4lIExF5U0R+EJGV7uRaMUtEbnb/P1kmItPdeWRihohMFZFMEVkWUNZMRD4WkdXuf5vWxmdZgogMRcAEVe0ODAZ+LyLdPY7Ja3/AGSbewBPAR6p6LNCHGG4XEWkL3AgMUNWegA9nrplYMg13orUAtwGfqmpn4FN3/bBZgogAqrpFVRe6y/txfgG09TYq77izCZ4NvOB1LF4TkcbAycCLAKpaoKp7vI3Kc/FAAxGJB5KBzR7HU6dU9QtgV6XiMcBL7vJLwHm18VmWICKMiHQE+gHfeBuJpx4HbgVKvA4kAnQCtgP/53a5vSAiMTsvrapuAh7GmY1yC7BXVed6G1VESFPVLe7yViCtNg5qCSKCiEgq8BZwk6ru8zoeL4jIOUCmqi7wOpYIEQ8cBzyrqv2AbGqp+yAauX3rY3AS5xFAiohc5m1UkcWdlbNW3l+wBBEhRCQBJzm8pqpvex2Ph4YAo0VkAzADOFVEXvU2JE9lABmqWnpF+SZOwohVpwPrVXW7qhYCbwMnehxTJNgmIm0A3P9m1sZBLUFEABERnD7mlar6qNfxeElVb1fVdqraEefm42eqGrN/IarqVmCjiHR1i04DVngYktd+BgaLSLL7/81pxPBN+wCzgF+7y78G3q2Ng1qCiAxDgMtx/lpe5P6M8jooEzFuAF4TkSVAX+A+j+PxjHsl9SawEFiK8zsspobdEJHpwDygq4hkiMjVwAPAGSKyGucq64Fa+SwbasMYY0wwdgVhjDEmKEsQxhhjgrIEYYwxJihLEMYYY4KyBGGMMSYoSxDGhJGIdAwcddOYaGIJwhhjTFCWIIypIyJylDvg3vFex2JMKOK9DsCYWOAOlTEDGKeqi72Ox5hQWIIwJvxa4oyN8wtVjeVxlEyUsS4mY8JvL84gc0O9DsSYQ2FXEMaEXwFwPjBHRLJU9XWvAzImFJYgjKkDqprtTob0sZskZnkdkzHVsdFcjTHGBGX3IIwxxgRlCcIYY0xQliCMMcYEZQnCGGNMUJYgjDHGBGUJwhhjTFCWIIwxxgT1/2qFrICTUJGfAAAAAElFTkSuQmCC\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 295
        },
        "id": "F4dBKkcKRcGG",
        "outputId": "03c825ff-1d79-48c5-bdf4-31cd0e405323"
      },
      "source": [
        "Accuracy_list = []\r\n",
        "for k in range(1, 11):\r\n",
        "  knn = MyKNeighborsClassifier(n_neighbours=k, weights=\"distance\", metric=\"manhattan\")\r\n",
        "  knn.fit(X_train, y_train)\r\n",
        "  y_pred = knn.predict(X_test)\r\n",
        "  Accuracy_list.append(MyAccuracyScore(y_test, y_pred))\r\n",
        "\r\n",
        "Accuracy_list_sklearn = []\r\n",
        "for k in range(1, 11):\r\n",
        "  knn_sklearn = KNeighborsClassifier(n_neighbors=k, weights='distance', metric='manhattan')\r\n",
        "  knn_sklearn.fit(X_train, y_train)\r\n",
        "  y_pred_sklearn = knn_sklearn.predict(X_test)\r\n",
        "  Accuracy_list_sklearn.append(accuracy_score(y_test, y_pred_sklearn))\r\n",
        "\r\n",
        "fig, ax = plt.subplots()\r\n",
        "ax.plot(range(1, 11), Accuracy_list, '-', label='My kNN', linewidth=2)\r\n",
        "ax.plot(range(1, 11), Accuracy_list_sklearn, '--', label='Sklearns kNN', linewidth=3)\r\n",
        "ax.legend()\r\n",
        "\r\n",
        "plt.title('Веса, зависящие от расстояния + Манхэттенское расстояние')\r\n",
        "plt.grid()\r\n",
        "plt.xlabel('k')\r\n",
        "plt.ylabel('Accuracy, %')\r\n",
        "plt.show()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "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": "markdown",
      "metadata": {
        "id": "FDMoI3t6b2Ea"
      },
      "source": [
        "3. Напишите вывод, сравнение всего, что получилось – получилось ли у вас достичь таких же результатов, как в sklearn, как на результат влияют параметры, какая модель и с какими параметрами оказалась лучшей.\r\n",
        "\r\n",
        "\r\n",
        "\r\n",
        "  Были достигнуты результаты работы модели, аналогичные с результатами работы встроенной библиотечной модели. В результате работы было выяснено, что при использовании классификатора sklearn, при одинаковом максимальном количестве соседей, приоритет отдается классу с наименьшим порядковым номером.\r\n",
        "\r\n",
        "  Например, если при k = 100 и 3 классах выяснилось, что было 40 соседей класса 0 и 40 соседей класса 1 (т.о., осталось 20 соседей класса 2), то модель определит это как класс 0, поскольку этот класс обладает меньшим порядковым номером ( 0 < 1).\r\n",
        "\r\n",
        "  Для представленного набора данных наиболее эффективным оказалось использование модели с реализацией манхэттенского расстояния при k = 5. В данном случае, доля правильных ответов составила 0.92.\r\n",
        "\r\n",
        "  Вторым по эффективности оказалось использование евклидова расстояния при k = 7. В данном случае, доля правильных ответов составила 0.91.\r\n",
        "\r\n",
        "  Наиболее неэффективным классификатором выступили все вариации параметров при k < 3.\r\n",
        "\r\n",
        "  Наиболее стабильные результаты показала модель с реализацией весов. при 7 из 10 значениях k модель показала одинаковые результаты, соответствующие доле правильных ответов в 0.88."
      ]
    }
  ]
}