From fd25d0b1c82485cbf4f7fdda7b31b376e8c70375 Mon Sep 17 00:00:00 2001 From: Ayush Agarwal <86561124+ayush-agarwal-0502@users.noreply.github.com> Date: Sat, 18 Jun 2022 23:06:17 +0530 Subject: [PATCH] Add files via upload --- anomaly_detector_ML_project.ipynb | 570 ++ predictive_maintenance.csv | 10001 ++++++++++++++++++++++++++++ 2 files changed, 10571 insertions(+) create mode 100644 anomaly_detector_ML_project.ipynb create mode 100644 predictive_maintenance.csv diff --git a/anomaly_detector_ML_project.ipynb b/anomaly_detector_ML_project.ipynb new file mode 100644 index 0000000..9d0533d --- /dev/null +++ b/anomaly_detector_ML_project.ipynb @@ -0,0 +1,570 @@ +{ + "nbformat": 4, + "nbformat_minor": 0, + "metadata": { + "colab": { + "name": "anomaly_detector_ML_project.ipynb", + "provenance": [], + "collapsed_sections": [] + }, + "kernelspec": { + "name": "python3", + "display_name": "Python 3" + }, + "language_info": { + "name": "python" + } + }, + "cells": [ + { + "cell_type": "code", + "source": [ + "# Project by Ayush Agarwal \n", + "# PROJECT - PREDICTIVE MAINTAINENCE ANOMALY PREDICTION \n", + "# IIT BHU Varanasi , Electronics Engineering Student \n", + "# ayush.agarwal.ece20@itbhu.ac.in\n", + "# https://www.linkedin.com/in/ayush-agarwal-261041215/\n", + "# https://github.com/ayush-agarwal-0502" + ], + "metadata": { + "id": "zvTPdwsFz7_X" + }, + "execution_count": 125, + "outputs": [] + }, + { + "cell_type": "code", + "execution_count": 126, + "metadata": { + "id": "8jHu8geimBPU" + }, + "outputs": [], + "source": [ + "import pandas as pd \n", + "import numpy as np \n", + "import matplotlib.pyplot as plt \n", + "import seaborn as sns" + ] + }, + { + "cell_type": "code", + "source": [ + "X = pd.read_csv(\"/content/predictive_maintenance.csv\")" + ], + "metadata": { + "id": "YWkZdBakmSJi" + }, + "execution_count": 127, + "outputs": [] + }, + { + "cell_type": "code", + "source": [ + "print(\"The data looks like :\")\n", + "print(X.head())\n", + "print(X.columns)\n", + "print(len(X.columns))" + ], + "metadata": { + "id": "GhMW478rmWl1", + "colab": { + "base_uri": "https://localhost:8080/" + }, + "outputId": "1a659a9b-4c9f-4094-a892-8217987942f8" + }, + "execution_count": 128, + "outputs": [ + { + "output_type": "stream", + "name": "stdout", + "text": [ + "The data looks like :\n", + " UDI Product ID Type Air temperature [K] Process temperature [K] \\\n", + "0 1 M14860 M 298.1 308.6 \n", + "1 2 L47181 L 298.2 308.7 \n", + "2 3 L47182 L 298.1 308.5 \n", + "3 4 L47183 L 298.2 308.6 \n", + "4 5 L47184 L 298.2 308.7 \n", + "\n", + " Rotational speed [rpm] Torque [Nm] Tool wear [min] Target Failure Type \n", + "0 1551 42.8 0 0 No Failure \n", + "1 1408 46.3 3 0 No Failure \n", + "2 1498 49.4 5 0 No Failure \n", + "3 1433 39.5 7 0 No Failure \n", + "4 1408 40.0 9 0 No Failure \n", + "Index(['UDI', 'Product ID', 'Type', 'Air temperature [K]',\n", + " 'Process temperature [K]', 'Rotational speed [rpm]', 'Torque [Nm]',\n", + " 'Tool wear [min]', 'Target', 'Failure Type'],\n", + " dtype='object')\n", + "10\n" + ] + } + ] + }, + { + "cell_type": "code", + "source": [ + "print(X['Failure Type'].value_counts())\n", + "print(len(X))" + ], + "metadata": { + "colab": { + "base_uri": "https://localhost:8080/" + }, + "id": "jdD1kuzunn3S", + "outputId": "dd24c07a-be14-4250-baf4-51cad12a1b09" + }, + "execution_count": 129, + "outputs": [ + { + "output_type": "stream", + "name": "stdout", + "text": [ + "No Failure 9652\n", + "Heat Dissipation Failure 112\n", + "Power Failure 95\n", + "Overstrain Failure 78\n", + "Tool Wear Failure 45\n", + "Random Failures 18\n", + "Name: Failure Type, dtype: int64\n", + "10000\n" + ] + } + ] + }, + { + "cell_type": "code", + "source": [ + "from sklearn.preprocessing import OrdinalEncoder\n", + "ordinal_encoder = OrdinalEncoder()\n", + "print(X.columns)\n", + "\n", + "\n", + "ordinal_cols = ['Type']\n", + "X_ordinal = X[ordinal_cols]\n", + "X_ordinal = ordinal_encoder.fit_transform(X_ordinal)\n", + "X_ordinal = pd.DataFrame(X_ordinal, columns = ordinal_cols)\n", + "print(X_ordinal.head())\n", + "X.drop(ordinal_cols,axis=1,inplace=True)\n", + "\n", + "X[ordinal_cols] = X_ordinal[ordinal_cols]\n", + "print(X.head())" + ], + "metadata": { + "colab": { + "base_uri": "https://localhost:8080/" + }, + "id": "zeigdmw3oUIZ", + "outputId": "c7ab7a89-c2df-43c4-b0d9-3b55c0265e9f" + }, + "execution_count": 130, + "outputs": [ + { + "output_type": "stream", + "name": "stdout", + "text": [ + "Index(['UDI', 'Product ID', 'Type', 'Air temperature [K]',\n", + " 'Process temperature [K]', 'Rotational speed [rpm]', 'Torque [Nm]',\n", + " 'Tool wear [min]', 'Target', 'Failure Type'],\n", + " dtype='object')\n", + " Type\n", + "0 2.0\n", + "1 1.0\n", + "2 1.0\n", + "3 1.0\n", + "4 1.0\n", + " UDI Product ID Air temperature [K] Process temperature [K] \\\n", + "0 1 M14860 298.1 308.6 \n", + "1 2 L47181 298.2 308.7 \n", + "2 3 L47182 298.1 308.5 \n", + "3 4 L47183 298.2 308.6 \n", + "4 5 L47184 298.2 308.7 \n", + "\n", + " Rotational speed [rpm] Torque [Nm] Tool wear [min] Target Failure Type \\\n", + "0 1551 42.8 0 0 No Failure \n", + "1 1408 46.3 3 0 No Failure \n", + "2 1498 49.4 5 0 No Failure \n", + "3 1433 39.5 7 0 No Failure \n", + "4 1408 40.0 9 0 No Failure \n", + "\n", + " Type \n", + "0 2.0 \n", + "1 1.0 \n", + "2 1.0 \n", + "3 1.0 \n", + "4 1.0 \n" + ] + } + ] + }, + { + "cell_type": "code", + "source": [ + "X.drop(['UDI','Product ID','Failure Type'],axis=1,inplace=True)" + ], + "metadata": { + "id": "5anM7Crdmn0x" + }, + "execution_count": 131, + "outputs": [] + }, + { + "cell_type": "code", + "source": [ + "y = X['Target']\n", + "X.drop(['Target'],axis=1,inplace=True)\n", + "print(X.head())\n", + "print(y.head())" + ], + "metadata": { + "colab": { + "base_uri": "https://localhost:8080/" + }, + "id": "koDmsZEoo4T1", + "outputId": "4131a4c2-3415-4695-a993-0198fe075a31" + }, + "execution_count": 132, + "outputs": [ + { + "output_type": "stream", + "name": "stdout", + "text": [ + " Air temperature [K] Process temperature [K] Rotational speed [rpm] \\\n", + "0 298.1 308.6 1551 \n", + "1 298.2 308.7 1408 \n", + "2 298.1 308.5 1498 \n", + "3 298.2 308.6 1433 \n", + "4 298.2 308.7 1408 \n", + "\n", + " Torque [Nm] Tool wear [min] Type \n", + "0 42.8 0 2.0 \n", + "1 46.3 3 1.0 \n", + "2 49.4 5 1.0 \n", + "3 39.5 7 1.0 \n", + "4 40.0 9 1.0 \n", + "0 0\n", + "1 0\n", + "2 0\n", + "3 0\n", + "4 0\n", + "Name: Target, dtype: int64\n" + ] + } + ] + }, + { + "cell_type": "code", + "source": [ + "# scaling X here \n", + "#Importing the ROBUST SCALER\n", + "from sklearn.preprocessing import RobustScaler\n", + "print(X.head())\n", + "# Using the robust scaler \n", + "robust_scaler_1 = RobustScaler()\n", + "X = robust_scaler_1.fit_transform(X)\n", + "X = pd.DataFrame(X)\n", + "\n", + "#After scaling data be like \n", + "print(X.head())" + ], + "metadata": { + "colab": { + "base_uri": "https://localhost:8080/" + }, + "id": "lmZuICtQpEeM", + "outputId": "9c49d714-f232-4a23-d47d-00945b2021a6" + }, + "execution_count": 133, + "outputs": [ + { + "output_type": "stream", + "name": "stdout", + "text": [ + " Air temperature [K] Process temperature [K] Rotational speed [rpm] \\\n", + "0 298.1 308.6 1551 \n", + "1 298.2 308.7 1408 \n", + "2 298.1 308.5 1498 \n", + "3 298.2 308.6 1433 \n", + "4 298.2 308.7 1408 \n", + "\n", + " Torque [Nm] Tool wear [min] Type \n", + "0 42.8 0 2.0 \n", + "1 46.3 3 1.0 \n", + "2 49.4 5 1.0 \n", + "3 39.5 7 1.0 \n", + "4 40.0 9 1.0 \n", + " 0 1 2 3 4 5\n", + "0 -0.62500 -0.652174 0.253968 0.198529 -0.990826 1.0\n", + "1 -0.59375 -0.608696 -0.502646 0.455882 -0.963303 0.0\n", + "2 -0.62500 -0.695652 -0.026455 0.683824 -0.944954 0.0\n", + "3 -0.59375 -0.652174 -0.370370 -0.044118 -0.926606 0.0\n", + "4 -0.59375 -0.608696 -0.502646 -0.007353 -0.908257 0.0\n" + ] + } + ] + }, + { + "cell_type": "code", + "source": [ + "from sklearn.model_selection import train_test_split\n", + "X_train, X_test, y_train, y_test = train_test_split(X, y, train_size=0.8, test_size=0.2,\n", + " random_state=0)" + ], + "metadata": { + "id": "a0mlJ4JnpDtQ" + }, + "execution_count": 134, + "outputs": [] + }, + { + "cell_type": "code", + "source": [ + "X_train['Target'] = y_train \n", + "\n", + "X_train_non_anomaly = X_train.loc[X_train['Target']==0]\n", + "X_train_anomaly = X_train.loc[X_train['Target']==1]\n", + "\n", + "X_train.drop(['Target'],axis=1,inplace=True)\n", + "X_train_non_anomaly.drop(['Target'],axis=1,inplace=True)\n", + "X_train_anomaly.drop(['Target'],axis=1,inplace=True)\n", + "\n", + "print(X_train_non_anomaly)" + ], + "metadata": { + "colab": { + "base_uri": "https://localhost:8080/" + }, + "id": "x5vMHYFBsfAR", + "outputId": "030b8953-8258-46d7-d038-d017ba90ad3f" + }, + "execution_count": 135, + "outputs": [ + { + "output_type": "stream", + "name": "stdout", + "text": [ + " 0 1 2 3 4 5\n", + "7389 -0.09375 0.217391 1.079365 -0.558824 0.146789 0.0\n", + "9275 -0.59375 -0.608696 0.539683 -0.786765 -0.559633 0.0\n", + "2995 0.12500 -0.130435 0.248677 -0.198529 0.366972 0.0\n", + "5316 1.09375 1.173913 0.984127 -0.514706 -0.844037 1.0\n", + "356 -0.78125 -0.739130 -0.555556 0.647059 -0.412844 1.0\n", + "... ... ... ... ... ... ...\n", + "9225 -0.65625 -0.434783 1.677249 -1.213235 0.110092 0.0\n", + "4859 1.12500 1.000000 0.169312 -0.110294 -0.009174 0.0\n", + "3264 0.37500 0.000000 -0.253968 0.294118 0.733945 -1.0\n", + "9845 -0.56250 -0.434783 -0.433862 0.536765 -0.688073 0.0\n", + "2732 -0.12500 -0.391304 -0.830688 1.272059 0.275229 0.0\n", + "\n", + "[7736 rows x 6 columns]\n" + ] + }, + { + "output_type": "stream", + "name": "stderr", + "text": [ + "/usr/local/lib/python3.7/dist-packages/pandas/core/frame.py:4913: SettingWithCopyWarning: \n", + "A value is trying to be set on a copy of a slice from a DataFrame\n", + "\n", + "See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy\n", + " errors=errors,\n" + ] + } + ] + }, + { + "cell_type": "code", + "source": [ + "model_mean = np.mean(X_train_non_anomaly, axis=0)\n", + "model_cov = np.cov(X_train_non_anomaly, rowvar=0)\n", + "print(model_mean)\n", + "print(model_cov)\n", + "# model_mean_anomaly = np.mean(X_train_anomaly, axis=0)\n", + "# model_cov_anomaly = np.cov(X_train_anomaly, rowvar=0)\n", + "# print(model_mean_anomaly)\n", + "# print(model_cov_anomaly)" + ], + "metadata": { + "colab": { + "base_uri": "https://localhost:8080/" + }, + "id": "Rexzs25Hpd-o", + "outputId": "79bed578-2ac4-4222-ad6e-840706f8a1e5" + }, + "execution_count": 136, + "outputs": [ + { + "output_type": "stream", + "name": "stdout", + "text": [ + "0 -0.044068\n", + "1 -0.050228\n", + "2 0.196877\n", + "3 -0.035302\n", + "4 -0.018685\n", + "5 0.201396\n", + "dtype: float64\n", + "[[ 0.39112389 0.35830612 0.01782564 -0.01225468 0.00502206 0.00695309]\n", + " [ 0.35830612 0.42348947 0.01420997 -0.00951814 0.00531966 0.00553597]\n", + " [ 0.01782564 0.01420997 0.78474163 -0.54727231 0.00797626 -0.0026121 ]\n", + " [-0.01225468 -0.00951814 -0.54727231 0.48442063 -0.00924472 0.00189366]\n", + " [ 0.00502206 0.00531966 0.00797626 -0.00924472 0.33473396 0.00455359]\n", + " [ 0.00695309 0.00553597 -0.0026121 0.00189366 0.00455359 0.36460568]]\n" + ] + } + ] + }, + { + "cell_type": "code", + "source": [ + "from scipy.stats import multivariate_normal\n", + "\n", + "preds_x_non_anomaly = multivariate_normal.pdf(X_train_non_anomaly, mean=model_mean, cov=model_cov*5)# - multivariate_normal.pdf(X_train_non_anomaly, mean=model_mean_anomaly, cov=model_cov_anomaly*5)\n", + "preds_x_anomaly = multivariate_normal.pdf(X_train_anomaly, mean=model_mean, cov=model_cov*5) #- multivariate_normal.pdf(X_train_anomaly, mean=model_mean_anomaly, cov=model_cov_anomaly*5)\n", + "\n", + "# X_train_anomaly\n", + "# print(preds_x_non_anomaly)\n", + "# print(\"-------------------\")\n", + "# print(preds_x_anomaly)\n", + "# print(\"------------\")\n", + "\n", + "sns.kdeplot(data=preds_x_non_anomaly, shade=True)\n", + "plt.show()\n", + "sns.kdeplot(data=preds_x_anomaly, shade=True)\n", + "plt.show()\n", + "sns.kdeplot(data=preds_x_non_anomaly, shade=True)\n", + "sns.kdeplot(data=preds_x_anomaly, shade=True)\n", + "plt.show()\n", + "\n", + "# sns.scatterplot(x=preds_x_non_anomaly, y=(preds_x_non_anomaly<100))\n", + "# sns.scatterplot(x=preds_x_anomaly, y=(preds_x_anomaly<100))\n", + "# plt.show()\n", + "\n", + "# sns.scatterplot(x=preds_x_anomaly, y=(preds_x_anomaly<100))\n", + "# sns.scatterplot(x=preds_x_non_anomaly, y=(preds_x_non_anomaly<100))\n", + "# plt.show()" + ], + "metadata": { + "colab": { + "base_uri": "https://localhost:8080/", + "height": 761 + }, + "id": "5iYti8fip6y-", + "outputId": "18bdb71f-b55f-47c2-bff8-49c85903763d" + }, + "execution_count": 143, + "outputs": [ + { + "output_type": "display_data", + "data": { + "text/plain": [ + "
" + ], + "image/png": "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\n" + }, + "metadata": { + "needs_background": "light" + } + }, + { + "output_type": "display_data", + "data": { + "text/plain": [ + "
" + ], + "image/png": "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\n" + }, + "metadata": { + "needs_background": "light" + } + }, + { + "output_type": "display_data", + "data": { + "text/plain": [ + "
" + ], + "image/png": "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\n" + }, + "metadata": { + "needs_background": "light" + } + } + ] + }, + { + "cell_type": "code", + "source": [ + "# threshold = 0.001\n", + "# threshold2 = 0.004\n", + "threshold_3 = 0.0008\n", + "preds_x_train = multivariate_normal.pdf(X_train, mean=model_mean, cov=model_cov*5)