Anomaly_detection_GMM.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "GMM.ipynb",
      "provenance": [],
      "authorship_tag": "ABX9TyPLYJVIzkQI0etBxg6OvHSd",
      "include_colab_link": true
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "accelerator": "GPU"
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "view-in-github",
        "colab_type": "text"
      },
      "source": [
        "<a href=\"https://colab.research.google.com/github/nlptrinh/CinnamonAI/blob/master/Anomaly_detection_GMM.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "gutif8l-OP-a",
        "colab_type": "code",
        "outputId": "651e9185-49e6-4df0-e1d0-87d7c16c4be7",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 54
        }
      },
      "source": [
        "from google.colab import drive\n",
        "drive.mount('/content/drive')"
      ],
      "execution_count": 52,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Drive already mounted at /content/drive; to attempt to forcibly remount, call drive.mount(\"/content/drive\", force_remount=True).\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "1Qet1sEpC1J-",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "import os\n",
        "os.chdir(\"/content/drive/My Drive/CinnamonAI\")"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "5Zhq3WIqOKwk",
        "colab_type": "code",
        "outputId": "6398eb74-e625-4789-e59f-ea00cab63b42",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        }
      },
      "source": [
        "!pwd"
      ],
      "execution_count": 54,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "/content/drive/My Drive/CinnamonAI\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "qhZLYPx0cwVD",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "%matplotlib inline\n",
        "import matplotlib as mpl\n",
        "import matplotlib.pyplot as plt\n",
        "from math import pi,e"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "WSp38zLjcwuJ",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "from scipy.io import loadmat\n",
        "from scipy.io.arff import loadarff\n",
        "import numpy as np\n",
        "import pandas as pd\n",
        "from collections import OrderedDict"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "1K8Dcl_Scw3f",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "from sklearn.preprocessing import StandardScaler, MinMaxScaler\n",
        "from sklearn.decomposition import PCA\n",
        "from sklearn.model_selection import StratifiedKFold, train_test_split, GridSearchCV\n",
        "from sklearn.metrics import precision_recall_curve, roc_curve, average_precision_score, auc, make_scorer\n",
        "from sklearn.metrics import f1_score,accuracy_score\n",
        "from sklearn.pipeline import make_pipeline\n",
        "from sklearn.externals.joblib import Parallel, delayed"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "wx776qilcw89",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "import seaborn as sns\n",
        "sns.set(font_scale=1.5)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "M_7YHX-hi-zN",
        "colab_type": "text"
      },
      "source": [
        "# **Using PCA for detection**"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Vb_eohwFju0D",
        "colab_type": "text"
      },
      "source": [
        "**Load data and inspect**"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "ZkQy5cx0cxDN",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "data = loadmat('data/cardio.mat')"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "51vlkTsEcxJr",
        "colab_type": "code",
        "outputId": "484e7e0f-c958-4f71-890c-ca6a4c5c46f0",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        }
      },
      "source": [
        "data.keys()"
      ],
      "execution_count": 60,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "dict_keys(['__header__', '__version__', '__globals__', 'X', 'y'])"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 60
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "5LTbiaxgcxNr",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "X = data['X']\n",
        "y = data['y']"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "xQlVf9jocxX0",
        "colab_type": "code",
        "outputId": "dbeaa819-5034-4e31-d6de-32bd53034813",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        }
      },
      "source": [
        "X.shape, y.shape"
      ],
      "execution_count": 62,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "((1831, 21), (1831, 1))"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 62
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "PIL8divjj7V9",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "y = y.squeeze()"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "X8eqG6xlkXAg",
        "colab_type": "text"
      },
      "source": [
        "**Scale data to have zero mean and unit standard deviation**"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "N9kWVW8XkAF8",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "scaler = StandardScaler()\n",
        "X = scaler.fit_transform(X)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Lghu93FYsSWb",
        "colab_type": "text"
      },
      "source": [
        "**Split into test and trainign data sets¶**"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "GoFi08hln6i9",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "X_train, X_test, y_train, y_test = train_test_split(X, y, test_size = 0.1)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "RgmeuS9Vn94c",
        "colab_type": "code",
        "outputId": "57e9bbd3-7626-4eaf-ab76-98b79a66c9bc",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        }
      },
      "source": [
        "len(X_train)"
      ],
      "execution_count": 66,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "1647"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 66
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "gwrJuNMGv0bU",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "def score(a, b):\n",
        "    \"\"\"Normalized differnece of squares between a and b.\"\"\"\n",
        "\n",
        "    loss = np.sum((a-b)**2, axis=1)\n",
        "    loss = (loss-loss.min())/(loss.max() - loss.min())\n",
        "    return loss"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "n_k4Plrov0mE",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "def fit(n, X):\n",
        "    pca = PCA(n_components=n)\n",
        "    X_pca = pca.fit_transform(X)\n",
        "    X_pca_inverse = pca.inverse_transform(X_pca)\n",
        "    y_score = score(X, X_pca_inverse)\n",
        "    return (n, y_score)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "lqN6Hs_Vv01t",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "search = Parallel(n_jobs=-1)(delayed(fit)(n, X_train) for n in range(2,X.shape[1]))"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "rQdhRLJqv0_6",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "scores =  np.array([(n, average_precision_score(y_train, y_score)) for (n, y_score) in search])"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "DYr2AFbrv1G0",
        "colab_type": "code",
        "outputId": "28a1cf2e-158e-46b0-8858-a34fe2d324a0",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        }
      },
      "source": [
        "best = max(scores, key=lambda x: x[1])\n",
        "best"
      ],
      "execution_count": 71,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array([2.        , 0.43997947])"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 71
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "SXGD01A-v1Mc",
        "colab_type": "code",
        "outputId": "f23e1add-b3ee-4fd1-e142-d18caf65317d",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 276
        }
      },
      "source": [
        "plt.scatter(scores[:, 0], scores[:, 1])\n",
        "pass"
      ],
      "execution_count": 72,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYsAAAEDCAYAAADEAyg+AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nhgUREQkxLIiISOj/AYGUO0AlGolWAAAAAElFTkSuQmCC\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "snKsYhgwv1ow",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "n = int(best[0])\n",
        "n, y_score = fit(n, X_test)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "o5DLTfCow87W",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        },
        "outputId": "3fde4600-05a3-4da9-e5d8-5ba8be226eca"
      },
      "source": [
        "average_precision_score(y_test, y_score)"
      ],
      "execution_count": 74,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "0.6526929839638869"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 74
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "jF0gDgWHw9HI",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "df = pd.DataFrame(dict(score=y_score, y=y_test))"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "IH0OtIrxw9TR",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 276
        },
        "outputId": "9a561269-bc3b-4800-f34d-46eb8ecfbf79"
      },
      "source": [
        "g = sns.swarmplot(data=df, x='y', y='score')\n",
        "g.set_xticklabels(['Inlier', 'Outliner'])\n",
        "g.set_xlabel('')\n",
        "pass"
      ],
      "execution_count": 76,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAZYAAAEDCAYAAAAWUyJmAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nQwAAAABJRU5ErkJggg==\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "YvA2-ZB8w9ag",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        ""
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "WyakapGJsCTd",
        "colab_type": "text"
      },
      "source": [
        "# **Visualizing data**"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "V7PZtuK9sATd",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "pca = PCA(n_components=2)\n",
        "X_pca = pca.fit_transform(X_train)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "zaoLBEe1sJAQ",
        "colab_type": "code",
        "outputId": "ec4c5da5-101b-4e43-e973-bb1a59891d14",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 293
        }
      },
      "source": [
        "plt.plot(X_pca[y_train == 0,0],X_pca[y_train==0,1],'b^')\n",
        "plt.plot(X_pca[y_train == 1,0],X_pca[y_train==1,1],'go')"
      ],
      "execution_count": 78,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "[<matplotlib.lines.Line2D at 0x7f64132cee80>]"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 78
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAY4AAAEDCAYAAAAiKuN6AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n5lXOvLVWAAAAAElFTkSuQmCC\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "8banj844sixp",
        "colab_type": "text"
      },
      "source": [
        "# **Gaussian Distribution**"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "ZjhBHFkwzeQE",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "data = loadmat('data/cardio.mat')"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "AerAV3Ldzgv6",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "X = data['X']\n",
        "y = data['y']"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "eUT70wxlzhLl",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "scaler = StandardScaler()"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "E1n-aY08zhQR",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "X = scaler.fit_transform(X)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "WyY2QFjNzhTx",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "X_train, X_test, y_train, y_test = train_test_split(X, y, test_size = 0.1)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "PuGmXFCHsnQj",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "def estimateGaussian(X):\n",
        "    mu = np.sum(X,axis = 0)/X.shape[0]\n",
        "    sigma2 = np.sum((X-mu)**2,axis = 0)/X.shape[0] \n",
        "    return mu,sigma2\n",
        "def multivariateGaussian(X,mu,sigma2):\n",
        "    sigma2 = np.diag(sigma2)\n",
        "    k = len(mu)\n",
        "    p = (2*pi)**(-k/2)*np.linalg.det(sigma2)**(-0.5)*np.exp(-0.5*((((X-mu))@np.linalg.inv(sigma2))@(X-mu).T))\n",
        "    return np.diag(p)\n",
        "def selectThresholds(y,p):\n",
        "    bestEpsilon = 0\n",
        "    bestF1 = 0\n",
        "    f1_score = 0\n",
        "    stepsize = (max(p)-min(p))/10000\n",
        "    for i in np.arange(min(p),max(p),stepsize):\n",
        "        prediction = (p < i).astype(int)\n",
        "        prediction = prediction.reshape(prediction.shape[0],1)\n",
        "        fp = np.sum((prediction == 1) * (y == 0))\n",
        "\n",
        "        fn = np.sum((prediction == 0) * (y == 1))\n",
        "        tp = np.sum((prediction == 1) * (y == 1))\n",
        "\n",
        "        prec = tp / (tp + fp)\n",
        "        rec = tp / (tp + fn)\n",
        "        \n",
        "        f1_score = 2 * prec * rec / (prec + rec)\n",
        "        if f1_score > bestF1:\n",
        "            bestF1 = f1_score\n",
        "            bestEpsilon = i\n",
        "    return bestEpsilon,bestF1"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "sbZJ-bZLs0cZ",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "X_normally = X_train[np.where(y_train == 0)[0]]\n",
        "y_normally = y_train[np.where(y_train == 0)[0]]\n",
        "X_anormally = X_train[np.where(y_train == 1)[0]]\n",
        "y_anormally = y_train[np.where(y_train == 1)[0]]"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "2yfw0YoZtpsX",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "mu,sigma2 = estimateGaussian(X_train)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "jkLVY9jgt2l4",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "p = multivariateGaussian(X_train,mu,sigma2)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "sCDZlekdt5H9",
        "colab_type": "code",
        "outputId": "9572b14a-eb9b-4a72-8205-3101eac85303",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 71
        }
      },
      "source": [
        "epsilon,F1= selectThresholds(y_train,p)\n",
        "epsilon,F1"
      ],
      "execution_count": 88,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "/usr/local/lib/python3.6/dist-packages/ipykernel_launcher.py:23: RuntimeWarning: invalid value encountered in long_scalars\n"
          ],
          "name": "stderr"
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(8.177831951996897e-14, 0.534453781512605)"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 88
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "9cAgTildu-yG",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "p_test = multivariateGaussian(X_test,mu,sigma2)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "5-qVgvNYvAqT",
        "colab_type": "code",
        "outputId": "723c9356-a8d3-4337-ecd0-cf9b61b9b4b9",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 374
        }
      },
      "source": [
        "pre = p_test < epsilon\n",
        "pre"
      ],
      "execution_count": 90,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array([False, False,  True, False, False, False, False, False, False,\n",
              "       False, False, False, False, False, False, False, False, False,\n",
              "       False,  True,  True, False, False,  True, False, False, False,\n",
              "       False, False, False, False, False, False, False,  True, False,\n",
              "       False, False, False,  True, False, False, False, False, False,\n",
              "       False, False, False, False, False, False,  True, False,  True,\n",
              "       False, False,  True, False, False, False, False, False,  True,\n",
              "       False,  True,  True, False, False, False, False, False,  True,\n",
              "        True, False, False, False, False, False,  True,  True, False,\n",
              "       False, False, False, False, False,  True,  True, False, False,\n",
              "       False,  True, False, False, False,  True, False, False, False,\n",
              "       False, False, False, False,  True, False, False, False,  True,\n",
              "       False,  True, False,  True,  True, False, False, False,  True,\n",
              "        True, False,  True, False,  True,  True, False,  True,  True,\n",
              "       False,  True,  True,  True, False, False, False,  True, False,\n",
              "       False,  True,  True,  True,  True, False, False, False, False,\n",
              "       False, False, False, False, False, False, False, False, False,\n",
              "        True, False, False, False, False, False, False,  True, False,\n",
              "        True, False,  True, False,  True, False, False, False, False,\n",
              "       False, False, False, False, False, False, False,  True, False,\n",
              "       False, False,  True, False])"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 90
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "f9mBYRIgvF4r",
        "colab_type": "code",
        "outputId": "9ee9b3cf-a302-4da0-9b01-9609c325458d",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        }
      },
      "source": [
        "from sklearn.metrics import f1_score\n",
        "f1_score(pre,y_test),accuracy_score(pre,y_test)"
      ],
      "execution_count": 91,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(0.44444444444444436, 0.8097826086956522)"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 91
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "xiS-tTF2x50w",
        "colab_type": "text"
      },
      "source": [
        "# **Gaussian Mixer Models**"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "bkYRZfR-x9XQ",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "from scipy.stats import multivariate_normal"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "u3Pb-A3cx9wg",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "class GMM:\n",
        "    def __init__(self, k, max_iter=5):\n",
        "        self.k = k\n",
        "        self.max_iter = int(max_iter)\n",
        "\n",
        "    def initialize(self, X):\n",
        "        self.shape = X.shape\n",
        "        self.n, self.m = self.shape\n",
        "\n",
        "        self.phi = np.full(shape=self.k, fill_value=1/self.k)\n",
        "        self.weights = np.full( shape=self.shape, fill_value=1/self.k)\n",
        "        \n",
        "        random_row = np.random.randint(low=0, high=self.n, size=self.k)\n",
        "        self.reg_cov = 1e-6*np.identity(self.m)\n",
        "        self.mu = [  X[row_index,:] for row_index in random_row ]\n",
        "        self.sigma = [ (np.cov(X.T)+self.reg_cov)  for _ in range(self.k) ] \n",
        "\n",
        "    def e_step(self, X):\n",
        "        # E-Step: update weights and phi holding mu and sigma constant\n",
        "        self.weights = self.predict_proba(X)\n",
        "        self.phi = self.weights.mean(axis=0)\n",
        "    \n",
        "    def m_step(self, X):\n",
        "        # M-Step: update mu and sigma holding phi and weights constant\n",
        "        for i in range(self.k):\n",
        "            weight = self.weights[:, [i]]\n",
        "            total_weight = weight.sum()\n",
        "            self.mu[i] = (X * weight).sum(axis=0) / total_weight\n",
        "            self.sigma[i] = np.cov(X.T, \n",
        "                aweights=(weight/total_weight).flatten(), \n",
        "                bias=True) + self.reg_cov\n",
        "\n",
        "    def fit(self, X):\n",
        "        self.initialize(X)\n",
        "        \n",
        "        for iteration in range(self.max_iter):\n",
        "            self.e_step(X)\n",
        "            self.m_step(X)\n",
        "            \n",
        "    def predict_proba(self, X):\n",
        "        likelihood = np.zeros( (self.n, self.k) )\n",
        "        for i in range(self.k):\n",
        "            distribution = multivariate_normal(\n",
        "                mean=self.mu[i], \n",
        "                cov=self.sigma[i])\n",
        "            likelihood[:,i] = distribution.pdf(X)\n",
        "#             likelihood[:,i] = multivariateGaussian(X,self.mu[i],self.sigma[i])\n",
        "        \n",
        "        numerator = likelihood * self.phi\n",
        "        denominator = numerator.sum(axis=1)[:, np.newaxis]\n",
        "        weights = numerator / denominator\n",
        "        return weights\n",
        "    \n",
        "    def predict(self, X):\n",
        "        weights = self.predict_proba(X)\n",
        "        return np.argmax(weights, axis=1)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "ek62FZQ2x93L",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "data = loadmat('data/cardio.mat')\n",
        "X = data['X']\n",
        "y = data['y']"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "qk-xn6qpx-B5",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "scaler = StandardScaler()\n",
        "X = scaler.fit_transform(X)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "9q-uifsqx-HC",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "X_train, X_test, y_train, y_test = train_test_split(X, y, test_size = 0.1)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "WWQFHHUxx98Y",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "X_normally = X_train[np.where(y_train == 0)[0]]\n",
        "y_normally = y_train[np.where(y_train == 0)[0]]\n",
        "X_anormally = X_train[np.where(y_train == 1)[0]]\n",
        "y_anormally = y_train[np.where(y_train == 1)[0]]"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "KV_UhL3pyMuH",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "gmm = GMM(k=1, max_iter=10)\n",
        "gmm.fit(X_train)\n",
        "pre = gmm.predict(X_train)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "9DcO8RN_yM_p",
        "colab_type": "text"
      },
      "source": [
        "# **GMM with Sklearn for anomaly detection**"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "_uz5YbfGyNXm",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "data = loadmat('data/cardio.mat')\n",
        "X = data['X']\n",
        "y = data['y']"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "zWIXXDEIyNUX",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "scaler = StandardScaler()\n",
        "X = scaler.fit_transform(X)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "EXxy2Gbv05c7",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "X_train, X_test, y_train, y_test = train_test_split(X, y, test_size = 0.1)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "T3nuiLIX05_z",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "X_normally = X_train[np.where(y_train == 0)[0]]\n",
        "y_normally = y_train[np.where(y_train == 0)[0]]\n",
        "X_anormally = X_train[np.where(y_train == 1)[0]]\n",
        "y_anormally = y_train[np.where(y_train == 1)[0]]"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "NCsHrRph08oN",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "from sklearn.mixture import GaussianMixture"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "zAMyplnV0-qy",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "gmm = GaussianMixture(n_components=1)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "U0F8aXFn0_6b",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 85
        },
        "outputId": "3a2d5a06-78b9-4e9d-88c6-c4829a71de31"
      },
      "source": [
        "gmm.fit(X_normally)"
      ],
      "execution_count": 105,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "GaussianMixture(covariance_type='full', init_params='kmeans', max_iter=100,\n",
              "                means_init=None, n_components=1, n_init=1, precisions_init=None,\n",
              "                random_state=None, reg_covar=1e-06, tol=0.001, verbose=0,\n",
              "                verbose_interval=10, warm_start=False, weights_init=None)"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 105
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "4RM76iO-1AdO",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "score = gmm.score_samples(X_train)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "TyPGHJnQ0_4-",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 293
        },
        "outputId": "2acd4723-c562-4d29-c6ab-344dacfaf073"
      },
      "source": [
        "plt.plot(score)"
      ],
      "execution_count": 107,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "[<matplotlib.lines.Line2D at 0x7f64131ef160>]"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 107
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAZ4AAAEDCAYAAAAFhGKSAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "q1utCJns1Fjb",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 293
        },
        "outputId": "3fa4d086-3ad8-4d6b-b294-073a59862d57"
      },
      "source": [
        "anomaly_score = score[np.where(y_train == 1)[0]]\n",
        "plt.plot(anomaly_score)"
      ],
      "execution_count": 108,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "[<matplotlib.lines.Line2D at 0x7f64131be358>]"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 108
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAZ4AAAEDCAYAAAAFhGKSAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nKvpBHQAAAABJRU5ErkJggg==\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "KaGxsy4Y1U8U",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 293
        },
        "outputId": "705b5124-d733-4dbd-8d7c-8ac5ccc8d3be"
      },
      "source": [
        "nomaly_score = score[np.where(y_train == 0)[0]]\n",
        "plt.plot(nomaly_score)"
      ],
      "execution_count": 109,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "[<matplotlib.lines.Line2D at 0x7f641318e240>]"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 109
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAZ0AAAEDCAYAAADus9mRAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "hcnrqyf-1Xuh",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 71
        },
        "outputId": "4393ee2d-96e3-466e-854c-f7573e6a111b"
      },
      "source": [
        "epsilon,F1= selectThresholds(y_train,score)\n",
        "epsilon,F1"
      ],
      "execution_count": 110,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "/usr/local/lib/python3.6/dist-packages/ipykernel_launcher.py:23: RuntimeWarning: invalid value encountered in long_scalars\n"
          ],
          "name": "stderr"
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(-26.127970275301777, 0.6989247311827957)"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 110
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "4Ipn1EcX1znX",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "test_score = gmm.score_samples(X_test)\n",
        "pre = (test_score < epsilon).astype(int)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "QYDEEcTM11GA",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        },
        "outputId": "eed465ba-0a0f-402d-86c6-9c2579b114f0"
      },
      "source": [
        "f1_score(pre,y_test)"
      ],
      "execution_count": 112,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "0.6666666666666666"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 112
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "IrymX_kk11E1",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        },
        "outputId": "e3eeb9da-77ca-4d93-f61e-d40f9d92ceb2"
      },
      "source": [
        "accuracy_score(pre,y_test)"
      ],
      "execution_count": 113,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "0.9293478260869565"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 113
        }
      ]
    }
  ]
}