Local_Global_Outliers.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "Local_Global_Outliers",
      "provenance": [],
      "authorship_tag": "ABX9TyNw0SFWfYqgn8/VyVTWEqxE",
      "include_colab_link": true
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "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/NeuralClassifier/OutlierDetect/blob/main/Local_Global_Outliers.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "CqcaF1XcIif0"
      },
      "outputs": [],
      "source": [
        "import numpy as np\n",
        "import matplotlib.pyplot as plt\n",
        "from sklearn.neighbors import LocalOutlierFactor\n",
        "import zipfile\n",
        "import os\n",
        "import shutil\n",
        "import pandas as pd\n",
        "from sklearn.neighbors import NearestNeighbors"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "!git clone https://github.com/NeuralClassifier/OutlierDetect"
      ],
      "metadata": {
        "id": "scrgkW04IqCq"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "def unzipFile(file_name, exctraction_place):\n",
        "  with zipfile.ZipFile(file_name,\"r\") as zip_ref:\n",
        "      zip_ref.extractall(exctraction_place)\n",
        "def removeFolder(file_path):\n",
        "  try:\n",
        "    shutil.rmtree(file_path)\n",
        "  except OSError as e:\n",
        "    print(\"Error: %s - %s.\" % (e.filename, e.strerror))"
      ],
      "metadata": {
        "id": "7smHaFFrJR65"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "unzipFile('50p.zip','/content/')"
      ],
      "metadata": {
        "id": "skjTzqIGLIF-"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "df = pd.read_excel('/content/50p/3/50p_m3_dim2.xlsx')\n",
        "df.drop(['Unnamed: 0'],axis = 1,inplace = True)\n",
        "df"
      ],
      "metadata": {
        "id": "DKeVmhXVK8ag"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "(max(X[:, 0])+1)*(-1)"
      ],
      "metadata": {
        "id": "xpf5hCo1bmuu"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "X = np.array(df.iloc[:,:-1])\n",
        "clf = LocalOutlierFactor(n_neighbors=20, contamination=0.1)\n",
        "# use fit_predict to compute the predicted labels of the training samples\n",
        "# (when LOF is used for outlier detection, the estimator has no predict,\n",
        "# decision_function and score_samples methods).\n",
        "y_pred = clf.fit_predict(X)\n",
        "#n_errors = (y_pred != ground_truth).sum()\n",
        "X_scores = clf.negative_outlier_factor_\n",
        "\n",
        "plt.title(\"Local Outlier Factor (LOF)\")\n",
        "plt.scatter(X[:, 0], X[:, 1], color=\"k\", s=3.0, label=\"Data points\")\n",
        "# plot circles with radius proportional to the outlier scores\n",
        "radius = (X_scores.max() - X_scores) / (X_scores.max() - X_scores.min())\n",
        "plt.scatter(X[:, 0],X[:, 1],s=1000 * radius,edgecolors=\"maroon\",facecolors=\"none\",label=\"Outlier scores\",)\n",
        "plt.axis(\"tight\")\n",
        "plt.xlim(((max(X[:, 0])+1)*(-1), max(X[:, 0])+1))\n",
        "plt.ylim(((max(X[:, 1])+1)*(-1), max(X[:, 1])+1))\n",
        "#plt.xlabel(\"prediction errors: %d\" % (n_errors))\n",
        "legend = plt.legend(loc=\"upper left\")\n",
        "legend.legendHandles[0]._sizes = [10]\n",
        "legend.legendHandles[1]._sizes = [20]\n",
        "plt.show()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 281
        },
        "id": "UJn9_eJhYc-r",
        "outputId": "bfd54eef-b258-45d2-e66f-49bd321d68e9"
      },
      "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": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "df_lofnorm=pd.concat([df,pd.DataFrame(radius,columns=['normalized_LOF'])],axis=1)\n",
        "df_lofnorm"
      ],
      "metadata": {
        "id": "daxKnSnBeQVm"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "df_lofnorm.sort_values([\"normalized_LOF\"], ascending=False)"
      ],
      "metadata": {
        "id": "OBi0AG1HvaAY"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "#LOF\n",
        "\n",
        "red = [30,50,70,90]\n",
        "p70 = [1,2,3,4]\n",
        "redPath = '/content/OutlierDetect/Data/reduced_synthetic_data/'\n",
        "\n",
        "#path_reduced = \"/content/\"+str(self.reduction)+'p'\n",
        "#os.mkdir(path_reduced)\n",
        "#path = \"/content/\"+str(self.reduction)+'p'+\"/outlierScore_perData/\"\n",
        "#os.mkdir(path)\n",
        "\n",
        "for folders in red:\n",
        "  subfold = redPath+str(folders)+'p/'\n",
        "  p70 = [1,2,3,4]\n",
        "  for subsub in p70:\n",
        "    subfold = redPath+str(folders)+'p/'\n",
        "    subfold = subfold+str(subsub)+'/'\n",
        "    path = subfold+'/LOF/'\n",
        "    os.mkdir(path)\n",
        "    dim = [2, 10, 20, 40, 80, 100]\n",
        "    for files in dim:\n",
        "      filename = subfold+str(folders)+'p'+'_m'+str(subsub)+'_dim'+str(files)+'.xlsx'\n",
        "      df = pd.read_excel(filename)\n",
        "      df.drop(['Unnamed: 0'],axis=1,inplace=True)\n",
        "      X = np.array(df.iloc[:,:-1])\n",
        "      clf = LocalOutlierFactor(n_neighbors=20, contamination=0.1)\n",
        "      y_pred = clf.fit_predict(X)\n",
        "      X_scores = clf.negative_outlier_factor_\n",
        "      #plt.title(\"Local Outlier Factor (LOF)\")\n",
        "      #plt.scatter(X[:, 0], X[:, 1], color=\"k\", s=3.0, label=\"Data points\")\n",
        "      # plot circles with radius proportional to the outlier scores\n",
        "      radius = (X_scores.max() - X_scores) / (X_scores.max() - X_scores.min())\n",
        "      #plt.scatter(X[:, 0],X[:, 1],s=1000 * radius,edgecolors=\"maroon\",facecolors=\"none\",label=\"Outlier scores\",)\n",
        "      #plt.axis(\"tight\")\n",
        "      #plt.xlim(((max(X[:, 0])+1)*(-1), max(X[:, 0])+1))\n",
        "      #plt.ylim(((max(X[:, 1])+1)*(-1), max(X[:, 1])+1))\n",
        "      #plt.xlabel(\"prediction errors: %d\" % (n_errors))\n",
        "      #legend = plt.legend(loc=\"upper left\")\n",
        "      #legend.legendHandles[0]._sizes = [10]\n",
        "      #legend.legendHandles[1]._sizes = [20]\n",
        "      #plt.show()\n",
        "      df_lofnorm=pd.concat([df,pd.DataFrame(radius,columns=['normalized_LOF'])],axis=1)\n",
        "      df_lofnorm.to_excel(path+str(folders)+'p'+'_m'+str(subsub)+'_dim'+str(files)+'_LOF.xlsx')"
      ],
      "metadata": {
        "id": "QWaXAsPTU1GE"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        ""
      ],
      "metadata": {
        "id": "PgD4WDbW3Je_"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "def replace_lbls(lbl, indx):\n",
        "  for idx in indx:\n",
        "    lbl[idx] = 1\n",
        "  return lbl"
      ],
      "metadata": {
        "id": "4XI_u42q5v0u"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        ""
      ],
      "metadata": {
        "id": "BjILIY-t5v9I"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "#KNN\n",
        "\n",
        "\n",
        "def replace_lbls(lbl, indx):\n",
        "  for idx in indx:\n",
        "    lbl[idx] = 1\n",
        "  return lbl\n",
        "\n",
        "red = [30,50,70,90]\n",
        "p70 = [1,2,3,4]\n",
        "redPath = '/content/OutlierDetect/Data/reduced_synthetic_data/'\n",
        "\n",
        "#path_reduced = \"/content/\"+str(self.reduction)+'p'\n",
        "#os.mkdir(path_reduced)\n",
        "#path = \"/content/\"+str(self.reduction)+'p'+\"/outlierScore_perData/\"\n",
        "#os.mkdir(path)\n",
        "\n",
        "for folders in red:\n",
        "  subfold = redPath+str(folders)+'p/'\n",
        "  p70 = [1,2,3,4]\n",
        "  for subsub in p70:\n",
        "    subfold = redPath+str(folders)+'p/'\n",
        "    subfold = subfold+str(subsub)+'/'\n",
        "    path = subfold+'/KNNOut/'\n",
        "    os.mkdir(path)\n",
        "    dim = [2, 10, 20, 40, 80, 100]\n",
        "    for files in dim:\n",
        "      filename = subfold+str(folders)+'p'+'_m'+str(subsub)+'_dim'+str(files)+'.xlsx'\n",
        "      df = pd.read_excel(filename)\n",
        "      df.drop(['Unnamed: 0'],axis=1,inplace=True)\n",
        "      X = np.array(df.iloc[:,:-1])\n",
        "\n",
        "      # instantiate model\n",
        "      nbrs = NearestNeighbors(n_neighbors = 10)\n",
        "      # fit model\n",
        "      nbrs.fit(X)\n",
        "      # distances and indexes of k-neighbors from model outputs\n",
        "      distances, indexes = nbrs.kneighbors(X)\n",
        "      lbl = np.zeros((len(X),), dtype=int)\n",
        "      #print(distances)\n",
        "      #print(list(indexes))\n",
        "  \n",
        "      # plot mean of k-distances of each observation\n",
        "      plt.plot(distances.mean(axis =1))\n",
        "      plt.show()\n",
        "      # visually determine cutoff values > 0.15\n",
        "      thres = float(input('Enter threshold: '))\n",
        "      outlier_index = np.where(distances.mean(axis = 1) > thres)\n",
        "      #print(outlier_index)\n",
        "      lbl = replace_lbls(lbl,outlier_index)\n",
        "\n",
        "      #clf = LocalOutlierFactor(n_neighbors=20, contamination=0.1)\n",
        "      #y_pred = clf.fit_predict(X)\n",
        "      #X_scores = clf.negative_outlier_factor_\n",
        "      #plt.title(\"Local Outlier Factor (LOF)\")\n",
        "      #plt.scatter(X[:, 0], X[:, 1], color=\"k\", s=3.0, label=\"Data points\")\n",
        "      # plot circles with radius proportional to the outlier scores\n",
        "      #radius = (X_scores.max() - X_scores) / (X_scores.max() - X_scores.min())\n",
        "      #plt.scatter(X[:, 0],X[:, 1],s=1000 * radius,edgecolors=\"maroon\",facecolors=\"none\",label=\"Outlier scores\",)\n",
        "      #plt.axis(\"tight\")\n",
        "      #plt.xlim(((max(X[:, 0])+1)*(-1), max(X[:, 0])+1))\n",
        "      #plt.ylim(((max(X[:, 1])+1)*(-1), max(X[:, 1])+1))\n",
        "      #plt.xlabel(\"prediction errors: %d\" % (n_errors))\n",
        "      #legend = plt.legend(loc=\"upper left\")\n",
        "      #legend.legendHandles[0]._sizes = [10]\n",
        "      #legend.legendHandles[1]._sizes = [20]\n",
        "      #plt.show()\n",
        "      df_lofnorm=pd.concat([df,pd.DataFrame(lbl,columns=['KNN_ouliers'])],axis=1)\n",
        "      df_lofnorm.to_excel(path+str(folders)+'p'+'_m'+str(subsub)+'_dim'+str(files)+'_KNNOut.xlsx')\n",
        "      #break\n",
        "    #break\n",
        "  #break"
      ],
      "metadata": {
        "id": "eKtstRrg3Jp0"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "df_lofnorm"
      ],
      "metadata": {
        "id": "o59MaNmf3Ju1"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "%rm -rf /content/OutlierDetect"
      ],
      "metadata": {
        "id": "g3jYpkUS8gbO"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "!zip -r /content/reduced_synthetic_data.zip /content/reduced_synthetic_data"
      ],
      "metadata": {
        "id": "1grhSpLWbgsg"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "!zip -r /content/reduced_synthetic_data.zip . i /content/OutlierDetect/Data/reduced_synthetic_data"
      ],
      "metadata": {
        "id": "Jk-BW1qEU16t"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "removeFolder('/content/70p')"
      ],
      "metadata": {
        "id": "YqY6lznIJZAe"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        ""
      ],
      "metadata": {
        "id": "hrrVUsWAKeZV"
      },
      "execution_count": null,
      "outputs": []
    }
  ]
}