ppn_sklearn.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "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": "3.7.4"
    },
    "colab": {
      "name": "ppn_sklearn.ipynb",
      "provenance": [],
      "collapsed_sections": [],
      "include_colab_link": true
    }
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "view-in-github",
        "colab_type": "text"
      },
      "source": [
        "<a href=\"https://colab.research.google.com/github/mashyko/deep-learning/blob/master/ppn_sklearn.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "BKG1jWoL18Sx",
        "colab_type": "text"
      },
      "source": [
        "Sklearn perceptron modelの使用"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "cvaF1xAj18S0",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 68
        },
        "outputId": "99b1e57a-08a6-4853-b72c-a13f00f596cf"
      },
      "source": [
        "from sklearn import datasets\n",
        "import numpy as np\n",
        "from sklearn.linear_model import Perceptron\n",
        "\n",
        "iris = datasets.load_iris()\n",
        "X = iris.data[:, [2, 3]]\n",
        "y = iris.target\n",
        "\n",
        "print('Class labels:', np.unique(y))\n",
        "\n",
        "ppn = Perceptron(max_iter=10000, tol=0.00001, random_state=1)\n",
        "#ppn = Perceptron(n_iter=400, eta0=0.00001, random_state=0)\n",
        "\n",
        "ppn.fit(X,y)\n",
        "y_pred = ppn.predict(X)\n",
        "\n",
        "print('Misclassified samples: %d' % (y != y_pred).sum())\n",
        "\n",
        "from sklearn.metrics import accuracy_score\n",
        "\n",
        "print('Accuracy: %.2f' % accuracy_score(y, y_pred))"
      ],
      "execution_count": 1,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Class labels: [0 1 2]\n",
            "Misclassified samples: 37\n",
            "Accuracy: 0.75\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Aj-qAMGV18S4",
        "colab_type": "text"
      },
      "source": [
        "ミニバッチ様式の導入"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "YAODg1mv18S5",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "\n",
        "#from sklearn.cross_validation import train_test_split\n",
        "from sklearn.model_selection import train_test_split\n",
        "\n",
        "X_train, X_test, y_train, y_test = train_test_split(\n",
        "    X, y, test_size=0.3, random_state=0)\n"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Dt4Zhra918S7",
        "colab_type": "text"
      },
      "source": [
        "データの標準化"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "mzAX0ktZ18S8",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "\n",
        "from sklearn.preprocessing import StandardScaler\n",
        "\n",
        "sc = StandardScaler()\n",
        "sc.fit(X_train)\n",
        "X_train_std = sc.transform(X_train)\n",
        "X_test_std = sc.transform(X_test)\n"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "CgYX_Rwj18S-",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 51
        },
        "outputId": "fdb93223-0196-46f6-925c-dab1acd4a4a2"
      },
      "source": [
        "from sklearn.linear_model import Perceptron\n",
        "\n",
        "ppn = Perceptron(max_iter=40, tol=0.1, random_state=0)\n",
        "ppn.fit(X_train_std, y_train)\n",
        "\n",
        "y_pred = ppn.predict(X_test_std)\n",
        "print('Misclassified samples: %d' % (y_test != y_pred).sum())\n",
        "\n",
        "from sklearn.metrics import accuracy_score\n",
        "\n",
        "print('Accuracy: %.2f' % accuracy_score(y_test, y_pred))\n"
      ],
      "execution_count": 4,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Misclassified samples: 15\n",
            "Accuracy: 0.67\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "sVDlFvo018TB",
        "colab_type": "text"
      },
      "source": [
        "plot_decision_regions関数の定義"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "875wFKZX18TB",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "from matplotlib.colors import ListedColormap\n",
        "import matplotlib.pyplot as plt\n",
        "\n",
        "def plot_decision_regions(X, y, classifier, test_idx=None, resolution=0.02):\n",
        "\n",
        "    # setup marker generator and color map\n",
        "    markers = ('s', 'x', 'o', '^', 'v')\n",
        "    colors = ('red', 'blue', 'lightgreen', 'gray', 'cyan')\n",
        "    cmap = ListedColormap(colors[:len(np.unique(y))])\n",
        "\n",
        "    # plot the decision surface\n",
        "    x1_min, x1_max = X[:, 0].min() - 1, X[:, 0].max() + 1\n",
        "    x2_min, x2_max = X[:, 1].min() - 1, X[:, 1].max() + 1\n",
        "    xx1, xx2 = np.meshgrid(np.arange(x1_min, x1_max, resolution),\n",
        "                           np.arange(x2_min, x2_max, resolution))\n",
        "    Z = classifier.predict(np.array([xx1.ravel(), xx2.ravel()]).T)\n",
        "    Z = Z.reshape(xx1.shape)\n",
        "    plt.contourf(xx1, xx2, Z, alpha=0.4, cmap=cmap)\n",
        "    plt.xlim(xx1.min(), xx1.max())\n",
        "    plt.ylim(xx2.min(), xx2.max())\n",
        "\n",
        "    for idx, cl in enumerate(np.unique(y)):\n",
        "        plt.scatter(x=X[y == cl, 0], \n",
        "                    y=X[y == cl, 1],\n",
        "                    alpha=0.6, \n",
        "                    c=cmap(idx),\n",
        "                    edgecolor='black',\n",
        "                    marker=markers[idx], \n",
        "                    label=cl)\n",
        "\n",
        "    # highlight test samples\n",
        "    if test_idx:\n",
        "        # plot all samples\n",
        "        X_test, y_test = X[test_idx, :], y[test_idx]\n",
        "        plt.scatter(X_test[:, 0],\n",
        "                    X_test[:, 1],\n",
        "                    c='',\n",
        "                    alpha=1.0,\n",
        "                    edgecolor='black',\n",
        "                    linewidths=1,\n",
        "                    marker='o',\n",
        "                    s=55, label='test set')\n"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "5x5qWjRS18TD",
        "colab_type": "text"
      },
      "source": [
        "データ分類の境界とその可視化"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "scrolled": true,
        "id": "jlOnzHDs18TE",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 368
        },
        "outputId": "3de63b73-c74a-42dd-a9b8-7a77998db1d1"
      },
      "source": [
        "X_combined_std = np.vstack((X_train_std, X_test_std))\n",
        "y_combined = np.hstack((y_train, y_test))\n",
        "\n",
        "plot_decision_regions(X=X_combined_std, y=y_combined,\n",
        "                      classifier=ppn, test_idx=range(105, 150))\n",
        "plt.xlabel('petal length [standardized]')\n",
        "plt.ylabel('petal width [standardized]')\n",
        "plt.legend(loc='upper left')\n",
        "\n",
        "plt.tight_layout()\n",
        "# plt.savefig('./figures/iris_perceptron_scikit.png', dpi=300)\n",
        "plt.show()"
      ],
      "execution_count": 6,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "'c' argument looks like a single numeric RGB or RGBA sequence, which should be avoided as value-mapping will have precedence in case its length matches with 'x' & 'y'.  Please use a 2-D array with a single row if you really want to specify the same RGB or RGBA value for all points.\n",
            "'c' argument looks like a single numeric RGB or RGBA sequence, which should be avoided as value-mapping will have precedence in case its length matches with 'x' & 'y'.  Please use a 2-D array with a single row if you really want to specify the same RGB or RGBA value for all points.\n",
            "'c' argument looks like a single numeric RGB or RGBA sequence, which should be avoided as value-mapping will have precedence in case its length matches with 'x' & 'y'.  Please use a 2-D array with a single row if you really want to specify the same RGB or RGBA value for all points.\n"
          ],
          "name": "stderr"
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAagAAAEYCAYAAAAJeGK1AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nrkJggg==\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "sAfbDGck18TG",
        "colab_type": "text"
      },
      "source": [
        "logistic regression model with scikit-learn"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "yh2IoPEl18TH",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 436
        },
        "outputId": "e3b72c13-d279-4f43-8ca4-e2d8de5f6a02"
      },
      "source": [
        "from sklearn.linear_model import LogisticRegression\n",
        "\n",
        "lr = LogisticRegression(C=1000.0, random_state=0)\n",
        "lr.fit(X_train_std, y_train)\n",
        "\n",
        "plot_decision_regions(X_combined_std, y_combined,\n",
        "                      classifier=lr, test_idx=range(105, 150))\n",
        "plt.xlabel('petal length [standardized]')\n",
        "plt.ylabel('petal width [standardized]')\n",
        "plt.legend(loc='upper left')\n",
        "plt.tight_layout()\n",
        "# plt.savefig('./figures/logistic_regression.png', dpi=300)\n",
        "plt.show()"
      ],
      "execution_count": 7,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "/usr/local/lib/python3.6/dist-packages/sklearn/linear_model/logistic.py:432: FutureWarning: Default solver will be changed to 'lbfgs' in 0.22. Specify a solver to silence this warning.\n",
            "  FutureWarning)\n",
            "/usr/local/lib/python3.6/dist-packages/sklearn/linear_model/logistic.py:469: FutureWarning: Default multi_class will be changed to 'auto' in 0.22. Specify the multi_class option to silence this warning.\n",
            "  \"this warning.\", FutureWarning)\n",
            "'c' argument looks like a single numeric RGB or RGBA sequence, which should be avoided as value-mapping will have precedence in case its length matches with 'x' & 'y'.  Please use a 2-D array with a single row if you really want to specify the same RGB or RGBA value for all points.\n",
            "'c' argument looks like a single numeric RGB or RGBA sequence, which should be avoided as value-mapping will have precedence in case its length matches with 'x' & 'y'.  Please use a 2-D array with a single row if you really want to specify the same RGB or RGBA value for all points.\n",
            "'c' argument looks like a single numeric RGB or RGBA sequence, which should be avoided as value-mapping will have precedence in case its length matches with 'x' & 'y'.  Please use a 2-D array with a single row if you really want to specify the same RGB or RGBA value for all points.\n"
          ],
          "name": "stderr"
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAagAAAEYCAYAAAAJeGK1AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nSUVORK5CYII=\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "6F3x2x7p18TJ",
        "colab_type": "text"
      },
      "source": [
        "MNIST データセット　with sklearn"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "vOQ368uw18TJ",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        },
        "outputId": "0a347af3-b820-482f-f9a4-58d0aa04b7b5"
      },
      "source": [
        "from sklearn import datasets, svm\n",
        "import numpy as np\n",
        "from sklearn.linear_model import Perceptron\n",
        "\n",
        "digits = datasets.load_digits()\n",
        "\n",
        "X = digits.data\n",
        "y = digits.target\n",
        "\n",
        "ppn = Perceptron(max_iter=10000, tol=0.00001, random_state=0)\n",
        "#ppn = Perceptron(n_iter=400, eta0=0.00001, random_state=0)\n",
        "\n",
        "ppn.fit(X,y)\n",
        "y_pred = ppn.predict(X)\n",
        "\n",
        "acc = sum(y_pred==y)/len(X)\n",
        "\n",
        "print('training accuracy = % .2f%%' % (acc * 100))\n",
        "\n"
      ],
      "execution_count": 8,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "training accuracy =  93.93%\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "RudkEjIu18TM",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        },
        "outputId": "e9040c94-2122-4cc5-af4a-aa7b10a6a61c"
      },
      "source": [
        "from sklearn import datasets, svm\n",
        "import numpy as np\n",
        "from sklearn.linear_model import Perceptron\n",
        "\n",
        "digits = datasets.load_digits()\n",
        "\n",
        "X = digits.data\n",
        "y = digits.target\n",
        "\n",
        "#from sklearn.cross_validation import train_test_split\n",
        "from sklearn.model_selection import train_test_split\n",
        "\n",
        "X_train, X_test, y_train, y_test = train_test_split(\n",
        "    X, y, test_size=0.3, random_state=0)\n",
        "\n",
        "ppn = Perceptron(max_iter=10000, tol=0.00001, random_state=0)\n",
        "#ppn = Perceptron(n_iter=400, eta0=0.00001, random_state=0)\n",
        "\n",
        "ppn.fit(X_train, y_train)\n",
        "y_test_pred = ppn.predict(X_test)\n",
        "\n",
        "acc = np.sum(y_test_pred==y_test, axis=0)/X_test.shape[0]\n",
        "\n",
        "\n",
        "print('Test data accuracy = % .2f%%' % (acc * 100))\n",
        "\n",
        "\n"
      ],
      "execution_count": 9,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Test data accuracy =  93.33%\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "WqHaX7G_18TN",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        ""
      ],
      "execution_count": 0,
      "outputs": []
    }
  ]
}