19AIE205_Python_for_Machine_Learning_Assignment_10.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "19AIE205 Python for Machine Learning-Assignment 10.ipynb",
      "provenance": [],
      "collapsed_sections": [],
      "authorship_tag": "ABX9TyM+m0jfVDafMAu+SQseJrZZ",
      "include_colab_link": true
    },
    "kernelspec": {
      "display_name": "Python 3",
      "name": "python3"
    },
    "language_info": {
      "name": "python"
    }
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "view-in-github",
        "colab_type": "text"
      },
      "source": [
        "<a href=\"https://colab.research.google.com/github/Jayku88/Machine-Learning-Lab/blob/main/19AIE205_Python_for_Machine_Learning_Assignment_10.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "EGnxUGtRBB-4"
      },
      "source": [
        "![MLoriginal.png](data:image/png;base64,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)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "YE82WMz9p2o1"
      },
      "source": [
        "# Topics\n",
        "1.   Support Vector Machine Classification (Implementation Only)\n",
        "2.   Parameter Tuning\n",
        "3. Visualization\n",
        "4. Importance of Preprocessing\n",
        "5. Advantages and Disadvantages of SVM\n",
        "6. Comparison of Different Classifiers\n",
        "\n",
        "\n",
        "\n",
        "\n",
        "\n",
        "\n",
        "\n"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "2m24ctauWE7v"
      },
      "source": [
        "#1.Support Vector Machine Classification (Implementation Only)"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "bLLWtY-T4t2I"
      },
      "source": [
        "import numpy as np\n",
        "import matplotlib.pyplot as plt\n",
        "from sklearn import svm, datasets\n",
        "iris = datasets.load_iris()\n",
        "X = iris.data[:, :2]\n",
        "y = iris.target\n",
        "svc = svm.SVC().fit(X, y) \n",
        "print(\"Accuracy: {:.3f}\".format(svc.score(X, y)))"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "RonObtAZIim2"
      },
      "source": [
        "#2.Parameter Tuning\n",
        "Important parameters having higher impact on model performance are **kernel, gamma and C.**\n",
        "\n",
        "**kernel**: we have various options available with kernel like, “linear”, “rbf”,”poly” and others (default value is “rbf”).  “rbf” and “poly” are useful for non-linear hyper-plane.\n",
        "\n",
        "**gamma**: Kernel coefficient for ‘rbf’, ‘poly’ and ‘sigmoid’. Higher the value of gamma, will try to exact fit the as per training data set, which will results in  generalization error and cause over-fitting problem.\n",
        "\n",
        "**C**: Penalty parameter C of the error term. It also controls the trade-off between smooth decision boundaries and classifying the training points correctly."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "XEAnm8UqK9C2"
      },
      "source": [
        "##kernel=\"linear\" and C=1"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "LM8uGQnVKDpJ"
      },
      "source": [
        "import numpy as np\n",
        "import matplotlib.pyplot as plt\n",
        "from sklearn import svm, datasets\n",
        "iris = datasets.load_iris()\n",
        "X = iris.data[:, :2]\n",
        "y = iris.target\n",
        "svc = svm.SVC(kernel=\"linear\").fit(X, y)\n",
        "print(\"Accuracy: {:.3f}\".format(svc.score(X, y)))"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "lnxdLtk3LGE2"
      },
      "source": [
        "##kernel=\"linear\" and C=100"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "T_i5VqdeK6Lw"
      },
      "source": [
        "import numpy as np\n",
        "import matplotlib.pyplot as plt\n",
        "from sklearn import svm, datasets\n",
        "iris = datasets.load_iris()\n",
        "X = iris.data[:, :2]\n",
        "y = iris.target\n",
        "svc = svm.SVC(kernel=\"linear\",C=100).fit(X, y)\n",
        "print(\"Accuracy: {:.3f}\".format(svc.score(X, y)))"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "GPUEIeTiLIQP"
      },
      "source": [
        "##kernel=\"linear\" and C=1000"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "nhkYYxoVK3TY"
      },
      "source": [
        "import numpy as np\n",
        "import matplotlib.pyplot as plt\n",
        "from sklearn import svm, datasets\n",
        "iris = datasets.load_iris()\n",
        "X = iris.data[:, :2]\n",
        "y = iris.target\n",
        "svc = svm.SVC(kernel=\"linear\",C=1000).fit(X, y)\n",
        "print(\"Accuracy: {:.3f}\".format(svc.score(X, y)))"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "_UWt-NqyMIMv"
      },
      "source": [
        "##kernel=\"rbf\", C=1, gamma=1"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "ecMT6pegLY1x"
      },
      "source": [
        "import numpy as np\n",
        "import matplotlib.pyplot as plt\n",
        "from sklearn import svm, datasets\n",
        "iris = datasets.load_iris()\n",
        "X = iris.data[:, :2]\n",
        "y = iris.target\n",
        "svc = svm.SVC(kernel=\"rbf\").fit(X, y)\n",
        "print(\"Accuracy: {:.3f}\".format(svc.score(X, y)))"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "OyK4AYlaMR7W"
      },
      "source": [
        "##kernel=\"rbf\", C=100, gamma=1"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "VaQWS3GaLY9j"
      },
      "source": [
        "import numpy as np\n",
        "import matplotlib.pyplot as plt\n",
        "from sklearn import svm, datasets\n",
        "iris = datasets.load_iris()\n",
        "X = iris.data[:, :2]\n",
        "y = iris.target\n",
        "svc = svm.SVC(kernel=\"rbf\",C=100).fit(X, y)\n",
        "print(\"Accuracy: {:.3f}\".format(svc.score(X, y)))"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "43kOMvvKMWfX"
      },
      "source": [
        "##kernel=\"rbf\", C=1000, gamma=1"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "w1abVy6PLZAV"
      },
      "source": [
        "import numpy as np\n",
        "import matplotlib.pyplot as plt\n",
        "from sklearn import svm, datasets\n",
        "iris = datasets.load_iris()\n",
        "X = iris.data[:, :2]\n",
        "y = iris.target\n",
        "svc = svm.SVC(kernel=\"rbf\",C=1000).fit(X, y)\n",
        "print(\"Accuracy: {:.3f}\".format(svc.score(X, y)))"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Uz7JVpyVMYbA"
      },
      "source": [
        "##kernel=\"rbf\", C=1, gamma=100"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "ZXOLycMxLZDJ"
      },
      "source": [
        "import numpy as np\n",
        "import matplotlib.pyplot as plt\n",
        "from sklearn import svm, datasets\n",
        "iris = datasets.load_iris()\n",
        "X = iris.data[:, :2]\n",
        "y = iris.target\n",
        "svc = svm.SVC(kernel=\"rbf\",gamma=100).fit(X, y)\n",
        "print(\"Accuracy: {:.3f}\".format(svc.score(X, y)))"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "dRS6NjeeMbMA"
      },
      "source": [
        "##kernel=\"rbf\", C=100, gamma=1000"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "oJx8xh0ILt7R"
      },
      "source": [
        "import numpy as np\n",
        "import matplotlib.pyplot as plt\n",
        "from sklearn import svm, datasets\n",
        "iris = datasets.load_iris()\n",
        "X = iris.data[:, :2]\n",
        "y = iris.target\n",
        "svc = svm.SVC(kernel=\"rbf\",gamma=1000).fit(X, y)\n",
        "print(\"Accuracy: {:.3f}\".format(svc.score(X, y)))"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "VFzsxhi5MwfC"
      },
      "source": [
        "#3.Visualization"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "hZv-VfqJNET6"
      },
      "source": [
        "##kernel=\"linear\" and C=1"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "VNhDVzZlLt_-"
      },
      "source": [
        "from sklearn.svm import SVC\n",
        "import numpy as np\n",
        "import matplotlib.pyplot as plt\n",
        "from sklearn import svm, datasets\n",
        "\n",
        "iris = datasets.load_iris()\n",
        "# Select 2 features / variable for the 2D plot that we are going to create.\n",
        "X = iris.data[:, :2]  # we only take the first two features.\n",
        "y = iris.target\n",
        "\n",
        "def make_meshgrid(x, y, h=.02):\n",
        "    x_min, x_max = x.min() - 1, x.max() + 1\n",
        "    y_min, y_max = y.min() - 1, y.max() + 1\n",
        "    xx, yy = np.meshgrid(np.arange(x_min, x_max, h), np.arange(y_min, y_max, h))\n",
        "    return xx, yy\n",
        "\n",
        "def plot_contours(ax, clf, xx, yy, **params):\n",
        "    Z = clf.predict(np.c_[xx.ravel(), yy.ravel()])\n",
        "    Z = Z.reshape(xx.shape)\n",
        "    out = ax.contourf(xx, yy, Z, **params)\n",
        "    return out\n",
        "\n",
        "model = svm.SVC(kernel='linear')\n",
        "clf = model.fit(X, y)\n",
        "\n",
        "fig, ax = plt.subplots()\n",
        "# title for the plots\n",
        "title = ('Decision surface of linear SVC, C=1 ')\n",
        "# Set-up grid for plotting.\n",
        "X0, X1 = X[:, 0], X[:, 1]\n",
        "xx, yy = make_meshgrid(X0, X1)\n",
        "\n",
        "plot_contours(ax, clf, xx, yy, cmap=plt.cm.coolwarm, alpha=0.8)\n",
        "ax.scatter(X0, X1, c=y, cmap=plt.cm.coolwarm, s=20, edgecolors='k')\n",
        "ax.set_xlabel('Sepal Length')\n",
        "ax.set_ylabel('Sepal Width')\n",
        "ax.set_xticks(())\n",
        "ax.set_yticks(())\n",
        "ax.set_title(title)\n",
        "plt.show()"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "7LnolRgGNhid"
      },
      "source": [
        "##kernel=\"linear\" and C=100"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "rBixo_KWM1vZ"
      },
      "source": [
        "from sklearn.svm import SVC\n",
        "import numpy as np\n",
        "import matplotlib.pyplot as plt\n",
        "from sklearn import svm, datasets\n",
        "\n",
        "iris = datasets.load_iris()\n",
        "# Select 2 features / variable for the 2D plot that we are going to create.\n",
        "X = iris.data[:, :2]  # we only take the first two features.\n",
        "y = iris.target\n",
        "\n",
        "def make_meshgrid(x, y, h=.02):\n",
        "    x_min, x_max = x.min() - 1, x.max() + 1\n",
        "    y_min, y_max = y.min() - 1, y.max() + 1\n",
        "    xx, yy = np.meshgrid(np.arange(x_min, x_max, h), np.arange(y_min, y_max, h))\n",
        "    return xx, yy\n",
        "\n",
        "def plot_contours(ax, clf, xx, yy, **params):\n",
        "    Z = clf.predict(np.c_[xx.ravel(), yy.ravel()])\n",
        "    Z = Z.reshape(xx.shape)\n",
        "    out = ax.contourf(xx, yy, Z, **params)\n",
        "    return out\n",
        "\n",
        "model = svm.SVC(kernel='linear',C=100)\n",
        "clf = model.fit(X, y)\n",
        "\n",
        "fig, ax = plt.subplots()\n",
        "# title for the plots\n",
        "title = ('Decision surface of linear SVC, C=100')\n",
        "# Set-up grid for plotting.\n",
        "X0, X1 = X[:, 0], X[:, 1]\n",
        "xx, yy = make_meshgrid(X0, X1)\n",
        "\n",
        "plot_contours(ax, clf, xx, yy, cmap=plt.cm.coolwarm, alpha=0.8)\n",
        "ax.scatter(X0, X1, c=y, cmap=plt.cm.coolwarm, s=20, edgecolors='k')\n",
        "ax.set_xlabel('Sepal Length')\n",
        "ax.set_ylabel('Sepal Width')\n",
        "ax.set_xticks(())\n",
        "ax.set_yticks(())\n",
        "ax.set_title(title)\n",
        "plt.show()"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "xgM7RffGNiDF"
      },
      "source": [
        "##kernel=\"linear\" and C=1000"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "s4-nLQNWNVa5"
      },
      "source": [
        "from sklearn.svm import SVC\n",
        "import numpy as np\n",
        "import matplotlib.pyplot as plt\n",
        "from sklearn import svm, datasets\n",
        "\n",
        "iris = datasets.load_iris()\n",
        "# Select 2 features / variable for the 2D plot that we are going to create.\n",
        "X = iris.data[:, :2]  # we only take the first two features.\n",
        "y = iris.target\n",
        "\n",
        "def make_meshgrid(x, y, h=.02):\n",
        "    x_min, x_max = x.min() - 1, x.max() + 1\n",
        "    y_min, y_max = y.min() - 1, y.max() + 1\n",
        "    xx, yy = np.meshgrid(np.arange(x_min, x_max, h), np.arange(y_min, y_max, h))\n",
        "    return xx, yy\n",
        "\n",
        "def plot_contours(ax, clf, xx, yy, **params):\n",
        "    Z = clf.predict(np.c_[xx.ravel(), yy.ravel()])\n",
        "    Z = Z.reshape(xx.shape)\n",
        "    out = ax.contourf(xx, yy, Z, **params)\n",
        "    return out\n",
        "\n",
        "model = svm.SVC(kernel='linear',C=1000)\n",
        "clf = model.fit(X, y)\n",
        "\n",
        "fig, ax = plt.subplots()\n",
        "# title for the plots\n",
        "title = ('Decision surface of linear SVC, C=1000')\n",
        "# Set-up grid for plotting.\n",
        "X0, X1 = X[:, 0], X[:, 1]\n",
        "xx, yy = make_meshgrid(X0, X1)\n",
        "\n",
        "plot_contours(ax, clf, xx, yy, cmap=plt.cm.coolwarm, alpha=0.8)\n",
        "ax.scatter(X0, X1, c=y, cmap=plt.cm.coolwarm, s=20, edgecolors='k')\n",
        "ax.set_xlabel('Sepal Length')\n",
        "ax.set_ylabel('Sepal Width')\n",
        "ax.set_xticks(())\n",
        "ax.set_yticks(())\n",
        "ax.set_title(title)\n",
        "plt.show()"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "VJMe2Yb7NxsZ"
      },
      "source": [
        "##kernel=\"rbf\", C=1, gamma=1"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "IFQpM1GnNVdG"
      },
      "source": [
        "from sklearn.svm import SVC\n",
        "import numpy as np\n",
        "import matplotlib.pyplot as plt\n",
        "from sklearn import svm, datasets\n",
        "\n",
        "iris = datasets.load_iris()\n",
        "# Select 2 features / variable for the 2D plot that we are going to create.\n",
        "X = iris.data[:, :2]  # we only take the first two features.\n",
        "y = iris.target\n",
        "\n",
        "def make_meshgrid(x, y, h=.02):\n",
        "    x_min, x_max = x.min() - 1, x.max() + 1\n",
        "    y_min, y_max = y.min() - 1, y.max() + 1\n",
        "    xx, yy = np.meshgrid(np.arange(x_min, x_max, h), np.arange(y_min, y_max, h))\n",
        "    return xx, yy\n",
        "\n",
        "def plot_contours(ax, clf, xx, yy, **params):\n",
        "    Z = clf.predict(np.c_[xx.ravel(), yy.ravel()])\n",
        "    Z = Z.reshape(xx.shape)\n",
        "    out = ax.contourf(xx, yy, Z, **params)\n",
        "    return out\n",
        "\n",
        "model = svm.SVC(kernel='rbf')\n",
        "clf = model.fit(X, y)\n",
        "\n",
        "fig, ax = plt.subplots()\n",
        "# title for the plots\n",
        "title = ('Decision surface of rbf SVC, C=1,gamma=1')\n",
        "# Set-up grid for plotting.\n",
        "X0, X1 = X[:, 0], X[:, 1]\n",
        "xx, yy = make_meshgrid(X0, X1)\n",
        "\n",
        "plot_contours(ax, clf, xx, yy, cmap=plt.cm.coolwarm, alpha=0.8)\n",
        "ax.scatter(X0, X1, c=y, cmap=plt.cm.coolwarm, s=20, edgecolors='k')\n",
        "ax.set_xlabel('Sepal Length')\n",
        "ax.set_ylabel('Sepal Width')\n",
        "ax.set_xticks(())\n",
        "ax.set_yticks(())\n",
        "ax.set_title(title)\n",
        "plt.show()"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "uz-uhlcJOQZF"
      },
      "source": [
        "##kernel=\"rbf\", C=100, gamma=1"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "y_HbmFizNVe9"
      },
      "source": [
        "from sklearn.svm import SVC\n",
        "import numpy as np\n",
        "import matplotlib.pyplot as plt\n",
        "from sklearn import svm, datasets\n",
        "\n",
        "iris = datasets.load_iris()\n",
        "# Select 2 features / variable for the 2D plot that we are going to create.\n",
        "X = iris.data[:, :2]  # we only take the first two features.\n",
        "y = iris.target\n",
        "\n",
        "def make_meshgrid(x, y, h=.02):\n",
        "    x_min, x_max = x.min() - 1, x.max() + 1\n",
        "    y_min, y_max = y.min() - 1, y.max() + 1\n",
        "    xx, yy = np.meshgrid(np.arange(x_min, x_max, h), np.arange(y_min, y_max, h))\n",
        "    return xx, yy\n",
        "\n",
        "def plot_contours(ax, clf, xx, yy, **params):\n",
        "    Z = clf.predict(np.c_[xx.ravel(), yy.ravel()])\n",
        "    Z = Z.reshape(xx.shape)\n",
        "    out = ax.contourf(xx, yy, Z, **params)\n",
        "    return out\n",
        "\n",
        "model = svm.SVC(kernel='rbf',C=100)\n",
        "clf = model.fit(X, y)\n",
        "\n",
        "fig, ax = plt.subplots()\n",
        "# title for the plots\n",
        "title = ('Decision surface of rbf SVC, C=100,gamma=1')\n",
        "# Set-up grid for plotting.\n",
        "X0, X1 = X[:, 0], X[:, 1]\n",
        "xx, yy = make_meshgrid(X0, X1)\n",
        "\n",
        "plot_contours(ax, clf, xx, yy, cmap=plt.cm.coolwarm, alpha=0.8)\n",
        "ax.scatter(X0, X1, c=y, cmap=plt.cm.coolwarm, s=20, edgecolors='k')\n",
        "ax.set_xlabel('Sepal Length')\n",
        "ax.set_ylabel('Sepal Width')\n",
        "ax.set_xticks(())\n",
        "ax.set_yticks(())\n",
        "ax.set_title(title)\n",
        "plt.show()"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "pfrxwqLBOr6d"
      },
      "source": [
        "##kernel=\"rbf\", C=1000, gamma=1"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "SafMeIsfOr6k"
      },
      "source": [
        "from sklearn.svm import SVC\n",
        "import numpy as np\n",
        "import matplotlib.pyplot as plt\n",
        "from sklearn import svm, datasets\n",
        "\n",
        "iris = datasets.load_iris()\n",
        "# Select 2 features / variable for the 2D plot that we are going to create.\n",
        "X = iris.data[:, :2]  # we only take the first two features.\n",
        "y = iris.target\n",
        "\n",
        "def make_meshgrid(x, y, h=.02):\n",
        "    x_min, x_max = x.min() - 1, x.max() + 1\n",
        "    y_min, y_max = y.min() - 1, y.max() + 1\n",
        "    xx, yy = np.meshgrid(np.arange(x_min, x_max, h), np.arange(y_min, y_max, h))\n",
        "    return xx, yy\n",
        "\n",
        "def plot_contours(ax, clf, xx, yy, **params):\n",
        "    Z = clf.predict(np.c_[xx.ravel(), yy.ravel()])\n",
        "    Z = Z.reshape(xx.shape)\n",
        "    out = ax.contourf(xx, yy, Z, **params)\n",
        "    return out\n",
        "\n",
        "model = svm.SVC(kernel='rbf',C=1000)\n",
        "clf = model.fit(X, y)\n",
        "\n",
        "fig, ax = plt.subplots()\n",
        "# title for the plots\n",
        "title = ('Decision surface of rbf SVC, C=1000,gamma=1')\n",
        "# Set-up grid for plotting.\n",
        "X0, X1 = X[:, 0], X[:, 1]\n",
        "xx, yy = make_meshgrid(X0, X1)\n",
        "\n",
        "plot_contours(ax, clf, xx, yy, cmap=plt.cm.coolwarm, alpha=0.8)\n",
        "ax.scatter(X0, X1, c=y, cmap=plt.cm.coolwarm, s=20, edgecolors='k')\n",
        "ax.set_xlabel('Sepal Length')\n",
        "ax.set_ylabel('Sepal Width')\n",
        "ax.set_xticks(())\n",
        "ax.set_yticks(())\n",
        "ax.set_title(title)\n",
        "plt.show()"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "_1g9ZeGzPOVK"
      },
      "source": [
        "##kernel=\"rbf\", C=1, gamma=100"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "97HRuk50POVU"
      },
      "source": [
        "from sklearn.svm import SVC\n",
        "import numpy as np\n",
        "import matplotlib.pyplot as plt\n",
        "from sklearn import svm, datasets\n",
        "\n",
        "iris = datasets.load_iris()\n",
        "# Select 2 features / variable for the 2D plot that we are going to create.\n",
        "X = iris.data[:, :2]  # we only take the first two features.\n",
        "y = iris.target\n",
        "\n",
        "def make_meshgrid(x, y, h=.02):\n",
        "    x_min, x_max = x.min() - 1, x.max() + 1\n",
        "    y_min, y_max = y.min() - 1, y.max() + 1\n",
        "    xx, yy = np.meshgrid(np.arange(x_min, x_max, h), np.arange(y_min, y_max, h))\n",
        "    return xx, yy\n",
        "\n",
        "def plot_contours(ax, clf, xx, yy, **params):\n",
        "    Z = clf.predict(np.c_[xx.ravel(), yy.ravel()])\n",
        "    Z = Z.reshape(xx.shape)\n",
        "    out = ax.contourf(xx, yy, Z, **params)\n",
        "    return out\n",
        "\n",
        "model = svm.SVC(kernel='rbf',gamma=100)\n",
        "clf = model.fit(X, y)\n",
        "\n",
        "fig, ax = plt.subplots()\n",
        "# title for the plots\n",
        "title = ('Decision surface of rbf SVC, C=1,gamma=100')\n",
        "# Set-up grid for plotting.\n",
        "X0, X1 = X[:, 0], X[:, 1]\n",
        "xx, yy = make_meshgrid(X0, X1)\n",
        "\n",
        "plot_contours(ax, clf, xx, yy, cmap=plt.cm.coolwarm, alpha=0.8)\n",
        "ax.scatter(X0, X1, c=y, cmap=plt.cm.coolwarm, s=20, edgecolors='k')\n",
        "ax.set_xlabel('Sepal Length')\n",
        "ax.set_ylabel('Sepal Width')\n",
        "ax.set_xticks(())\n",
        "ax.set_yticks(())\n",
        "ax.set_title(title)\n",
        "plt.show()"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "ZB8XEMkbPdUK"
      },
      "source": [
        "##kernel=\"rbf\", C=1, gamma=1000"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "w6rdPUKHPdUM"
      },
      "source": [
        "from sklearn.svm import SVC\n",
        "import numpy as np\n",
        "import matplotlib.pyplot as plt\n",
        "from sklearn import svm, datasets\n",
        "\n",
        "iris = datasets.load_iris()\n",
        "# Select 2 features / variable for the 2D plot that we are going to create.\n",
        "X = iris.data[:, :2]  # we only take the first two features.\n",
        "y = iris.target\n",
        "\n",
        "def make_meshgrid(x, y, h=.02):\n",
        "    x_min, x_max = x.min() - 1, x.max() + 1\n",
        "    y_min, y_max = y.min() - 1, y.max() + 1\n",
        "    xx, yy = np.meshgrid(np.arange(x_min, x_max, h), np.arange(y_min, y_max, h))\n",
        "    return xx, yy\n",
        "\n",
        "def plot_contours(ax, clf, xx, yy, **params):\n",
        "    Z = clf.predict(np.c_[xx.ravel(), yy.ravel()])\n",
        "    Z = Z.reshape(xx.shape)\n",
        "    out = ax.contourf(xx, yy, Z, **params)\n",
        "    return out\n",
        "\n",
        "model = svm.SVC(kernel='rbf',gamma=1000)\n",
        "clf = model.fit(X, y)\n",
        "\n",
        "fig, ax = plt.subplots()\n",
        "# title for the plots\n",
        "title = ('Decision surface of rbf SVC, C=1,gamma=1000')\n",
        "# Set-up grid for plotting.\n",
        "X0, X1 = X[:, 0], X[:, 1]\n",
        "xx, yy = make_meshgrid(X0, X1)\n",
        "\n",
        "plot_contours(ax, clf, xx, yy, cmap=plt.cm.coolwarm, alpha=0.8)\n",
        "ax.scatter(X0, X1, c=y, cmap=plt.cm.coolwarm, s=20, edgecolors='k')\n",
        "ax.set_xlabel('Sepal Length')\n",
        "ax.set_ylabel('Sepal Width')\n",
        "ax.set_xticks(())\n",
        "ax.set_yticks(())\n",
        "ax.set_title(title)\n",
        "plt.show()"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "3eTab6ZRPxBe"
      },
      "source": [
        "#4.Importance of Preprocessing"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "XlG90AwBbodq"
      },
      "source": [
        "## Before Preprocessing"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "fxFAGqk8euKz"
      },
      "source": [
        "from sklearn.svm import SVC\n",
        "from sklearn.model_selection import train_test_split\n",
        "from sklearn.datasets import load_breast_cancer\n",
        "cancer = load_breast_cancer()\n",
        "X_train, X_test, y_train, y_test = train_test_split(cancer.data, cancer.target)\n",
        "svm = SVC(C=1)\n",
        "svm.fit(X_train, y_train)\n",
        "print(\"Test set accuracy: {:.2f}\".format(svm.score(X_test, y_test)))\n",
        "print(\"Train set accuracy: {:.2f}\".format(svm.score(X_train, y_train)))"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "p5NvWZe3bxgC"
      },
      "source": [
        "## Result after Scaling with MinMaxScaler"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "OjO_nzdueuPK"
      },
      "source": [
        "from sklearn.preprocessing import MinMaxScaler\n",
        "scaler = MinMaxScaler()\n",
        "scaler.fit(X_train)\n",
        "X_train_scaled = scaler.transform(X_train)\n",
        "X_test_scaled = scaler.transform(X_test)\n",
        "# learning an SVM on the scaled training data\n",
        "svm.fit(X_train_scaled, y_train)\n",
        "# scoring on the scaled test set\n",
        "print(\"Scaled train set accuracy: {:.2f}\".format(svm.score(X_train_scaled, y_train)))\n",
        "print(\"Scaled test set accuracy: {:.2f}\".format(svm.score(X_test_scaled, y_test)))"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Lux9I8jbb9i8"
      },
      "source": [
        "## Result after Scaling with StandardScaler"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "UcP4wNFfeuUl"
      },
      "source": [
        "from sklearn.preprocessing import StandardScaler\n",
        "scaler = StandardScaler()\n",
        "scaler.fit(X_train)\n",
        "X_train_scaled = scaler.transform(X_train)\n",
        "X_test_scaled = scaler.transform(X_test)\n",
        "# learning an SVM on the scaled training data\n",
        "svm.fit(X_train_scaled, y_train)\n",
        "# scoring on the scaled test set\n",
        "print(\"Scaled train set accuracy: {:.2f}\".format(svm.score(X_train_scaled, y_train)))\n",
        "print(\"Scaled test set accuracy: {:.2f}\".format(svm.score(X_test_scaled, y_test)))"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "YqAWve2AcBKu"
      },
      "source": [
        "## Result after Scaling with RobustScaler"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "ANoxzAsAeBZe"
      },
      "source": [
        "from sklearn.preprocessing import RobustScaler\n",
        "scaler = RobustScaler()\n",
        "scaler.fit(X_train)\n",
        "X_train_scaled = scaler.transform(X_train)\n",
        "X_test_scaled = scaler.transform(X_test)\n",
        "# learning an SVM on the scaled training data\n",
        "svm.fit(X_train_scaled, y_train)\n",
        "# scoring on the scaled test set\n",
        "print(\"Scaled train set accuracy: {:.2f}\".format(svm.score(X_train_scaled, y_train)))\n",
        "print(\"Scaled test set accuracy: {:.2f}\".format(svm.score(X_test_scaled, y_test)))"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "8ohaPdYfcLrh"
      },
      "source": [
        "Comparison of different preprocessing techniques\n",
        "\n",
        "https://stackoverflow.com/questions/30918781/right-function-for-normalizing-input-of-sklearn-svm\n"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "_Wxt_S6jc_Fn"
      },
      "source": [
        "#5. Advantages and Disadvantages of SVM\n",
        "\n"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "NqseSpngfcuW"
      },
      "source": [
        "**Advantages**\n",
        "\n",
        "1. Effective in high dimensional spaces.\n",
        "\n",
        "2. Still effective in cases where number of dimensions is greater than the number of samples.\n",
        "\n",
        "3. Uses a subset of training points in the decision function (called support vectors), so it is also memory efficient.\n",
        "\n",
        "4. Versatile: different Kernel functions can be specified for the decision function. Common kernels are provided, but it is also possible to specify custom kernels.\n",
        "\n",
        "**Disadvantages**\n",
        "\n",
        "1. If the number of features is much greater than the number of samples, avoid over-fitting in choosing Kernel functions and regularization term is crucial.\n",
        "\n",
        "2. SVMs do not directly provide probability estimates, these are calculated using an expensive five-fold cross-validation."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "x_t2dz-EdY1L"
      },
      "source": [
        "#6. Comparsion of different Classifiers\n"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "C23M_LuYdzb0"
      },
      "source": [
        "**Nearest neighbors**\n",
        "\n",
        "For small datasets, good as a baseline, easy to explain.\n",
        "\n",
        "**Linear models**\n",
        "\n",
        "Go-to as a first algorithm to try, good for very large datasets, good for very highdimensional data.\n",
        "\n",
        "**Naive Bayes**\n",
        "\n",
        "Only for classification. Even faster than linear models, good for very large datasets and high-dimensional data. Often less accurate than linear models.\n",
        "\n",
        "**Decision trees**\n",
        "\n",
        "Very fast, don’t need scaling of the data, can be visualized and easily explained.\n",
        "\n",
        "**Random forests**\n",
        "\n",
        "Nearly always perform better than a single decision tree, very robust and powerful. Don’t need scaling of data. Not good for very high-dimensional sparse data.\n",
        "\n",
        "**Gradient boosted decision trees**\n",
        "\n",
        "Often slightly more accurate than random forests. Slower to train but faster to predict than random forests, and smaller in memory. Need more parameter tuning than random forests.\n",
        "\n",
        "**Support vector machines**\n",
        "\n",
        "Powerful for medium-sized datasets of features with similar meaning. Require\n",
        "scaling of data, sensitive to parameters.\n",
        "\n",
        "**Neural networks**\n",
        "\n",
        "Can build very complex models, particularly for large datasets. Sensitive to scaling of the data and to the choice of parameters. Large models need a long time to\n",
        "train."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "D9uSzNY283D7"
      },
      "source": [
        "# Reflections\n",
        "\n",
        "1.   Support Vector Machine Classification (Implementation Only)\n",
        "2.   Parameter Tuning\n",
        "3. Visualization\n",
        "4. Importance of Preprocessing\n",
        "5. Advantages and Disadvantages of SVM\n",
        "6. Comparison of Different Classifiers\n",
        "\n",
        "\n",
        "\n"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "cellView": "form",
        "id": "zlfSj5pyrGjD"
      },
      "source": [
        "#@title  { run: \"auto\", vertical-output: true }\n",
        "#@title  { run: \"auto\", vertical-output: true }\n",
        "name = \"\" #@param {type:\"string\"}\n",
        "roll_number = \"\" #@param {type:\"string\"}\n",
        "self_assessment = 10 #@param {type:\"slider\", min:1, max:10, step:1}\n",
        "print(\"Name : \",name)\n",
        "print(\"Roll Number : \",roll_number)\n",
        "print(\"Self Assessment Marks : \",self_assessment,\"/10\")"
      ],
      "execution_count": null,
      "outputs": []
    }
  ]
}