{

 "cells": [

  {

   "cell_type": "markdown",

   "id": "d298c50e",

   "metadata": {},

   "source": [

    "# Civil Structure Fault Classification using Vibration Signals"

   ]

  },

  {

   "cell_type": "markdown",

   "id": "17a02c7b",

   "metadata": {},

   "source": [

    "## Motivation"

   ]

  },

  {

   "cell_type": "markdown",

   "id": "25179f33",

   "metadata": {},

   "source": [

    "Thousands of lives are lost due to faults in civil structures each year. Detecting potential faults in large civil structures is key to preventing such mass calamities. Structural health monitoring (SHM) is an important and growing field with widespread application in civil engineering and architecture design. Structural health monitoring (SHM) will play a crucial role in making structures safer by detecting potential failures well before they actual happen.\n",

    "\n",

    "In this project, we want to create a model that will be able to predict the types of fault in a civil structure."

   ]

  },

  {

   "cell_type": "markdown",

   "id": "793ef327",

   "metadata": {},

   "source": [

    "## Dataset"

   ]

  },

  {

   "cell_type": "markdown",

   "id": "c5f71dc3",

   "metadata": {},

   "source": [

    "There are civil structure design benchmarks available for Structural Health Monitoring (SHM) prupose. One such benchmark model was established in the Earth-quake Engineering Research Facility at the University of BritishColumbia (UBC, Canada). The model used is a four-storey with two-bay-by-two-bay steel frame scale structure. The size of each plane is 1.5 m × 1.5 m, the height of each floor is 0.9 m, and each floor has eight braces. The benchmark model is shown in the following figure.\n"

   ]

  },

  {

   "cell_type": "markdown",

   "id": "ebac2b42",

   "metadata": {},

   "source": [

    "Civil structure layout used for simulation"

   ]

  },

  {

   "cell_type": "markdown",

   "id": "c3ac6644",

   "metadata": {},

   "source": [

    "![Experimental setup](image.png)"

   ]

  },

  {

   "cell_type": "markdown",

   "id": "e4ad17d5",

   "metadata": {},

   "source": [

    "Sensor placement setup"

   ]

  },

  {

   "cell_type": "markdown",

   "id": "b1af31f2",

   "metadata": {},

   "source": [

    "![Sensor setup](image1.PNG)"

   ]

  },

  {

   "cell_type": "markdown",

   "id": "71655795",

   "metadata": {},

   "source": [

    "#### Experimental conditions"

   ]

  },

  {

   "cell_type": "markdown",

   "id": "941049b0",

   "metadata": {},

   "source": [

    "A 140 data points - dataset is recorded with the help of 16 sensors. Each datapoint represents a vibration signal of length 40000. These signals are categorized in 6 types of fault and a normal class. In total, 7 classes that are used for classification by Machine learning models.\n",

    "\n",

    "The layout shown above was used to generate data in simulation software. The simulation software used for generating the data is **DataGen2e**. The software has a set of libraries and civil benchmark structures. After selecting the structure layout, we need to create conditions under which we want to create the data. The data for this project is collected with the following conditions.\n",

    "\n",

    "1. Undamaged case\n",

    "2. All braces of 1st story are broken\n",

    "3. all braces of 1st and 3rd story are broken\n",

    "4. 1 brace on 1st story is broken\n",

    "5. 1 brace on 3rd story is broken\n",

    "6. condition 5 + unscrew the element 18\n",

    "7. Area of brace 1 is reduced to 2/3rd of story 1\n",

    "\n",

    "After selecting the conditions, the structure is then excited using twice the natural frequency for 40 seconds under each condition. The vibration signals at each sensor are collected with 1000 Hz frequency. We have collected 20 datapoints for each condition.\n",

    "\n",

    "\n"

   ]

  },

  {

   "cell_type": "markdown",

   "id": "7e67d4fb",

   "metadata": {},

   "source": [

    "## Steps"

   ]

  },

  {

   "cell_type": "markdown",

   "id": "190044d7",

   "metadata": {},

   "source": [

    "**1. Data Processing and Visualization**\n",

    "\n",

    "**2. Feature Engineering**\n",

    "\n",

    "**3. Machine Learning methods and performace evaluation**\n",

    "\n",

    "**4. Deep Learning method and performace evaluation**\n",

    "\n",

    "**5. Conclusion**\n",

    "\n",

    "\n",

    "**Appendix**"

   ]

  },

  {

   "cell_type": "markdown",

   "id": "84f8b5b6",

   "metadata": {},

   "source": [

    "We will import all the required libraries here"

   ]

  },

  {

   "cell_type": "code",

   "execution_count": 24,

   "id": "c3a52379",

   "metadata": {},

   "outputs": [],

   "source": [

    "# Regular libraries\n",

    "import numpy as np\n",

    "import pandas as pd\n",

    "import scipy.stats\n",

    "import os\n",

    "\n",

    "# Graphing libraries\n",

    "import matplotlib.pyplot as plt\n",

    "import seaborn as sn\n",

    "\n",

    "# Machine Learning libraries\n",

    "from sklearn.ensemble import RandomForestClassifier\n",

    "from sklearn.svm import SVC\n",

    "from sklearn.linear_model import LogisticRegression\n",

    "from sklearn.neighbors import KNeighborsClassifier\n",

    "\n",

    "# Model evaluation libraries\n",

    "from sklearn.model_selection import cross_val_score\n",

    "from sklearn.metrics import accuracy_score\n",

    "from sklearn.metrics import confusion_matrix\n",

    "\n",

    "# Warning handling library\n",

    "import warnings\n",

    "warnings.filterwarnings('ignore')"

   ]

  },

  {

   "cell_type": "markdown",

   "id": "918460b1",

   "metadata": {},

   "source": [

    "## 1. Data Processing and Visualization"

   ]

  },

  {

   "cell_type": "markdown",

   "id": "7400ab43",

   "metadata": {},

   "source": [

    "Let us now load the data. The data is stored as an array in '.npy' file format. Labels are also stored in the same file format. After loading the data, we will print the shape and visualize the data."

   ]

  },

  {

   "cell_type": "code",

   "execution_count": 25,

   "id": "1ee26af0",

   "metadata": {},

   "outputs": [],

   "source": [

    "data = np.load('data.npy', allow_pickle = True)\n",

    "labels = np.load('labels.npy', allow_pickle = True)"

   ]

  },

  {

   "cell_type": "code",

   "execution_count": 26,

   "id": "986b5711",

   "metadata": {},

   "outputs": [

    {

     "name": "stdout",

     "output_type": "stream",

     "text": [

      "Number of data points:  140\n",

      "Number of sensors:  16\n",

      "Signal length:  40000\n",

      "Classes:  [0 1 2 3 4 5 6]\n"

     ]

    }

   ],

   "source": [

    "print('Number of data points: ',data.shape[0])\n",

    "print('Number of sensors: ',data.shape[1])\n",

    "print('Signal length: ',data.shape[2])\n",

    "print('Classes: ', np.unique(labels))"

   ]

  },

  {

   "cell_type": "markdown",

   "id": "cbbd8b28",

   "metadata": {},

   "source": [

    "The data loaded above is in raw format. We will process the data as per the requirement of models we will be using.\n",

    "\n",

    "Let us now plot the signal captured by different sensors."

   ]

  },

  {

   "cell_type": "code",

   "execution_count": 27,

   "id": "c866d903",

   "metadata": {},

   "outputs": [

    {

     "data": {

      "text/plain": [

       "[Text(0, 0.5, 'Amplitude'), Text(0.5, 0, 'Time')]"

      ]

     },

     "execution_count": 27,

     "metadata": {},

     "output_type": "execute_result"

    },

    {

     "data": {

      "image/png": "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\n",

      "text/plain": [

       "<Figure size 1080x576 with 4 Axes>"

      ]

     },

     "metadata": {},

     "output_type": "display_data"

    }

   ],

   "source": [

    "fig, axs = plt.subplots(nrows=2, ncols=2, figsize=(15, 8))\n",

    "fig.subplots_adjust(hspace=.35)\n",

    "((ax1, ax2), (ax3, ax4)) = axs\n",

    "plt.subplots_adjust(bottom=0.1, right=0.8, top=0.9)\n",

    "ax1.plot(data[0, 0, :], label = 'Vibration signal')\n",

    "ax1.set_title('Sensor 1')\n",

    "ax1.set(ylabel='Amplitude', xlabel='Time')\n",

    "ax2.plot(data[0, 4, :], label = 'Vibration signal')\n",

    "ax2.set_title('Sensor 4')\n",

    "ax2.set(ylabel='Amplitude', xlabel='Time')\n",

    "ax3.plot(data[0, 8, :], label = 'Vibration signal')\n",

    "ax3.set_title('Sensor 8')\n",

    "ax3.set(ylabel='Amplitude', xlabel='Time')\n",

    "ax4.plot(data[0, 12, :], label = 'Vibration signal')\n",

    "ax4.set_title('Sensor 12')\n",

    "ax4.set(ylabel='Amplitude', xlabel='Time')"

   ]

  },

  {

   "cell_type": "markdown",

   "id": "b359b90d",

   "metadata": {},

   "source": [

    "These signals are very much alike. Vibration signals are generally normalized before we start feature extraction. The normalization of the signal comes under preprocessing step. There are a few other steps as well that are performed during preprocessing such as noise removal, outlier analysis. For the purpose of this project, we will only normalize the signals."

   ]

  },

  {

   "cell_type": "markdown",

   "id": "c8e34e06",

   "metadata": {},

   "source": [

    "Normalization is a rescaling of the data from the original range so that all values are within the range of 0 and 1.\n",

    "Normalization can be useful, and even required in some machine learning algorithms when your time series data has input values with differing scales."

   ]

  },

  {

   "cell_type": "markdown",

   "id": "89f2b19e",

   "metadata": {},

   "source": [

    "There are different normalization techniques such as RMS normalization, mean normalization. Here we will be using Min-Max scaling method."

   ]

  },

  {

   "cell_type": "code",

   "execution_count": 28,

   "id": "68f82f0e",

   "metadata": {},

   "outputs": [],

   "source": [

    "data = (data - np.min(data, axis = 2, keepdims = True))/(np.max(data, axis = 2, keepdims = True) - \\\n",

    "                                                                    np.min(data, axis = 2, keepdims = True))"

   ]

  },

  {

   "cell_type": "markdown",

   "id": "a4d4255e",

   "metadata": {},

   "source": [

    "### Scaled data after Normalization"

   ]

  },

  {

   "cell_type": "code",

   "execution_count": 29,

   "id": "0addd20f",

   "metadata": {},

   "outputs": [

    {

     "data": {

      "text/plain": [

       "[Text(0, 0.5, 'Amplitude'), Text(0.5, 0, 'Time')]"

      ]

     },

     "execution_count": 29,

     "metadata": {},

     "output_type": "execute_result"

    },

    {

     "data": {

      "image/png": "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\n",

      "text/plain": [

       "<Figure size 1080x576 with 4 Axes>"

      ]

     },

     "metadata": {},

     "output_type": "display_data"

    }

   ],

   "source": [

    "fig, axs = plt.subplots(nrows=2, ncols=2, figsize=(15, 8))\n",

    "fig.subplots_adjust(hspace=.35)\n",

    "((ax1, ax2), (ax3, ax4)) = axs\n",

    "plt.subplots_adjust(bottom=0.1, right=0.8, top=0.9)\n",

    "ax1.plot(data[0, 0, :], label = 'Vibration signal')\n",

    "ax1.set_title('Sensor 1')\n",

    "ax1.set(ylabel='Amplitude', xlabel='Time')\n",

    "ax2.plot(data[0, 4, :], label = 'Vibration signal')\n",

    "ax2.set_title('Sensor 4')\n",

    "ax2.set(ylabel='Amplitude', xlabel='Time')\n",

    "ax3.plot(data[0, 8, :], label = 'Vibration signal')\n",

    "ax3.set_title('Sensor 8')\n",

    "ax3.set(ylabel='Amplitude', xlabel='Time')\n",

    "ax4.plot(data[0, 12, :], label = 'Vibration signal')\n",

    "ax4.set_title('Sensor 12')\n",

    "ax4.set(ylabel='Amplitude', xlabel='Time')"

   ]

  },

  {

   "cell_type": "markdown",

   "id": "2725d4d2",

   "metadata": {},

   "source": [

    "Let us save the processed data that we will use to train the models later."

   ]

  },

  {

   "cell_type": "code",

   "execution_count": 30,

   "id": "5ff37db6",

   "metadata": {},

   "outputs": [],

   "source": [

    "np.save('data_processed.npy', data)\n",

    "np.save('labels_processed.npy', labels)"

   ]

  },

  {

   "cell_type": "markdown",

   "id": "0be91573",

   "metadata": {},

   "source": [

    "## 2. Feature Engineering"

   ]

  },

  {

   "cell_type": "markdown",

   "id": "7653ec36",

   "metadata": {},

   "source": [

    "To train our machine learning models, we will use Feature Engineering. Before that, let us see what Feature Engineering is."

   ]

  },

  {

   "cell_type": "markdown",

   "id": "8ef425e6",

   "metadata": {},

   "source": [

    "### What is Feature Engineering?"

   ]

  },

  {

   "cell_type": "markdown",

   "id": "72bd09c4",

   "metadata": {},

   "source": [

    "Feature Engineering is an important step in solving Machine Learning problem. It is a process to extract useful information based on domain knowledge from raw data. The extracted useful information is a set of features that help machine learning algorithm to solve the problem with improved performance. "

   ]

  },

  {

   "cell_type": "markdown",

   "id": "a05c364a",

   "metadata": {},

   "source": [

    "A feature is a numeric representation of raw data. There are many ways to extract features from raw data. The right features are relevant to the task at hand and should be easy for the model to ingest. Feature engineering is the process of formulating the most appropriate features given the data, the model, and the task. "

   ]

  },

  {

   "cell_type": "markdown",

   "id": "0b7c9356",

   "metadata": {},

   "source": [

    "The number of features is also important. If there are not enough informative features, then the model will be unable to perform the ultimate task. If there are too many features, or if most of them are irrelevant, then the model will be more expensive and tricky to train. Something might go awry in the training process that impacts the model’s performance."

   ]

  },

  {

   "cell_type": "markdown",

   "id": "549bff69",

   "metadata": {},

   "source": [

    "### Feature Extraction"

   ]

  },

  {

   "cell_type": "markdown",

   "id": "7304958d",

   "metadata": {},

   "source": [

    "After normalizing the signal, we will start extracting features to help shallow Machine Learning Algorithms perform better on this dataset. We will be extracting simple yet powerful statistical features to classify these signals. There are few features that are specific to vibration signal analysis such as Crest factor, Shape factor. At the root, these features are also statistical in nature.\n",

    "\n",

    "1. mean\n",

    "2. median\n",

    "3. min\n",

    "4. max\n",

    "5. peak_to_peak\n",

    "6. variance\n",

    "7. rms\n",

    "8. absolute_mean\n",

    "9. shape_factor\n",

    "10. impulse_factor\n",

    "11. crest_factor\n",

    "12. clearance_factor\n",

    "13. std\n",

    "14. skewness\n",

    "15. kurtosis\n",

    "16. abslogmean\n",

    "17. meanabsdev\n",

    "18. medianabsdev\n",

    "19. midrange\n",

    "20. coeff_var\n",

    "\n",

    "In the appendix, all the features are explained."

   ]

  },

  {

   "cell_type": "markdown",

   "id": "8533eeb3",

   "metadata": {},

   "source": [

    "The are a few reasons behind extracting these features. These features are easy-to-calculate. Besides, these features captures the signal characteristics good enough in most of the cases. The paper published by W. Caesarendra (attached in Reference section, number 10) discusses the benefits of using these features over using complicated features such as frequency domain features. In some cases, the frequency domain features reveal more information about the underlying nature of the signals. In most of the cases, these features are sufficient."

   ]

  },

  {

   "cell_type": "markdown",

   "id": "6811a42b",

   "metadata": {},

   "source": [

    "Following class comprises of 20 features."

   ]

  },

  {

   "cell_type": "code",

   "execution_count": 1,

   "id": "29401d92",

   "metadata": {},

   "outputs": [],

   "source": [

    "class Featurizer():\n",

    "    \n",

    "    def __init__(self, data, axis = 1):\n",

    "        self.data = data\n",

    "        self.axis = axis\n",

    "    \n",

    "    def mean(self):\n",

    "        ans = np.mean(self.data, self.axis)\n",

    "        return ans\n",

    "\n",

    "    def median(self):\n",

    "        ans = np.median(self.data, self.axis)\n",

    "        return ans\n",

    "\n",

    "    def min_value(self):\n",

    "        ans = np.min(self.data, self.axis)\n",

    "        return ans\n",

    "\n",

    "    def max_value(self):\n",

    "        ans = np.max(self.data, self.axis)\n",

    "        return ans\n",

    "\n",

    "    def peak_to_peak(self):\n",

    "        ans = np.max(self.data, self.axis) - np.min(self.data, self.axis)\n",

    "        return ans\n",

    "\n",

    "    def variance(self):\n",

    "        ans = np.var(self.data, self.axis)\n",

    "        return ans\n",

    "\n",

    "    def rms(self):\n",

    "        ans = np.sqrt(np.mean(self.data ** 2, self.axis))\n",

    "        return ans\n",

    "\n",

    "    def abs_mean(self):\n",

    "        ans = np.mean(np.absolute(self.data), self.axis)\n",

    "        return ans\n",

    "\n",

    "    def shapefactor(self):\n",

    "        ans = self.rms() / self.abs_mean()\n",

    "        return ans\n",

    "\n",

    "    def impulsefactor(self):\n",

    "        ans = np.max(np.absolute(self.data), self.axis) / self.abs_mean()\n",

    "        return ans\n",

    "\n",

    "    def crestfactor(self):\n",

    "        ans = np.max(np.absolute(self.data), self.axis) / np.sqrt(np.mean(self.data ** 2, self.axis))\n",

    "        return ans\n",

    "\n",

    "    def clearancefactor(self):\n",

    "        ans = np.max(np.absolute(self.data), self.axis)\n",

    "        ans /= ((np.mean(np.sqrt(np.absolute(self.data)), self.axis)) ** 2)\n",

    "        return ans\n",

    "\n",

    "    def std(self):\n",

    "        ans = np.std(self.data, self.axis)\n",

    "        return ans\n",

    "\n",

    "    def skew(self):\n",

    "        ans = scipy.stats.skew(self.data, self.axis)\n",

    "        return ans\n",

    "\n",

    "    def kurtosis(self):\n",

    "        ans = scipy.stats.kurtosis(self.data, self.axis)\n",

    "        return ans\n",

    "\n",

    "    def abslogmean(self):\n",

    "        ans = np.mean(np.log(np.abs(self.data)+1e-12), self.axis)\n",

    "        return ans\n",

    "\n",

    "    def meanabsdev(self):\n",

    "        if self.axis == 0:\n",

    "            ans = np.mean(np.abs(self.data - np.mean(self.data, self.axis)), self.axis)\n",

    "        else:\n",

    "            ans = np.mean(\n",

    "                np.abs(self.data - np.mean(self.data, self.axis).reshape(self.data.shape[0], 1)), self.axis)\n",

    "        return ans\n",

    "\n",

    "    def medianabsdev(self):\n",

    "        if self.axis == 0:\n",

    "            ans = np.median(np.abs(self.data - np.median(self.data, self.axis)), self.axis)\n",

    "        else:\n",

    "            ans = np.median(\n",

    "                np.abs(self.data - np.median(self.data, self.axis).reshape(self.data.shape[0], 1)), self.axis)\n",

    "        return ans\n",

    "\n",

    "    def midrange(self):\n",

    "        ans = (np.max(self.data, self.axis) + np.min(self.data, self.axis)) / 2\n",

    "        return ans\n",

    "\n",

    "    def coeff_var(self):\n",

    "        ans = scipy.stats.variation(self.data, self.axis)\n",

    "        return ans\n",

    "    \n",

    "    all_funcs = [mean, median, min_value, max_value, peak_to_peak, variance, \\\n",

    "                rms, abs_mean, shapefactor, impulsefactor,crestfactor, clearancefactor, \\\n",

    "                std, skew, kurtosis, abslogmean, meanabsdev, medianabsdev, midrange, coeff_var]\n",

    "    \n",

    "    features = ['mean', 'median', 'min_value', 'max_value', 'peak_to_peak', 'variance', \\\n",

    "                'rms', 'abs_mean', 'shapefactor', 'impulsefactor', 'crestfactor', 'clearancefactor', \\\n",

    "                'std', 'skew', 'kurtosis', 'abslogmean', 'meanabsdev', 'medianabsdev', 'midrange', 'coeff_var']\n",

    "        "

   ]

  },

  {

   "cell_type": "markdown",

   "id": "95bbc577",

   "metadata": {},

   "source": [

    "Let us now use the above class to extract the features for our dataset. The class needs to be instantiated with a dataset and axis along which we want to extract features. We will extract features for each sensor's output"

   ]

  },

  {

   "cell_type": "markdown",

   "id": "b6487e79",

   "metadata": {},

   "source": [

    "Loading the data"

   ]

  },

  {

   "cell_type": "code",

   "execution_count": 32,

   "id": "41e8fafa",

   "metadata": {},

   "outputs": [],

   "source": [

    "data = np.load('data.npy', allow_pickle = True)"

   ]

  },

  {

   "cell_type": "code",

   "execution_count": 33,

   "id": "2993025e",

   "metadata": {},

   "outputs": [],

   "source": [

    "num_sensors = 16\n",

    "featurized_data = []\n",

    "for sensor in range(num_sensors):\n",

    "    sensor_feature = []\n",

    "    f = Featurizer(data[:, sensor, :], axis = 1)\n",

    "    for func in f.all_funcs:\n",

    "        sensor_feature.append(func(f))\n",

    "    featurized_data.append(np.array(sensor_feature).T)\n",

    "featurized_data = np.array(featurized_data)"

   ]

  },

  {

   "cell_type": "code",

   "execution_count": 34,

   "id": "1d305881",

   "metadata": {},

   "outputs": [

    {

     "name": "stdout",

     "output_type": "stream",

     "text": [

      "Shape of the featurized data:  (16, 140, 20)\n"

     ]

    }

   ],

   "source": [

    "print('Shape of the featurized data: ', featurized_data.shape)"

   ]

  },

  {

   "cell_type": "markdown",

   "id": "7f76f47a",

   "metadata": {},

   "source": [

    "Now we will merge the features from all the signals of each datapoint to form a complete feature vector"

   ]

  },

  {

   "cell_type": "code",

   "execution_count": 35,

   "id": "2b243641",

   "metadata": {},

   "outputs": [],

   "source": [

    "num_datapoints = featurized_data.shape[1]\n",

    "final_data = []\n",

    "for i in range(num_datapoints):\n",

    "    final_data.append(featurized_data[:, i, :].ravel())\n",

    "final_data = np.array(final_data)"

   ]

  },

  {

   "cell_type": "code",

   "execution_count": 36,

   "id": "c246ea8c",

   "metadata": {},

   "outputs": [

    {

     "name": "stdout",

     "output_type": "stream",

     "text": [

      "Final shape of the featurized data:  (140, 320)\n"

     ]

    }

   ],

   "source": [

    "print('Final shape of the featurized data: ', final_data.shape)"

   ]

  },

  {

   "cell_type": "markdown",

   "id": "ab1280a6",

   "metadata": {},

   "source": [

    "We have 16 sensors. 20 features are extracted from each sensor's output. As a result we have 20*16 = 320 features per datapoint."

   ]

  },

  {

   "cell_type": "markdown",

   "id": "6e553f83",

   "metadata": {},

   "source": [

    "Saving the featurized data"

   ]

  },

  {

   "cell_type": "code",

   "execution_count": 37,

   "id": "aacae2e2",

   "metadata": {},

   "outputs": [],

   "source": [

    "np.save('featurized_data.npy', final_data)"

   ]

  },

  {

   "cell_type": "markdown",

   "id": "0f2547bb",

   "metadata": {},

   "source": [

    " Now its time to train models and see how our feature extraction works.\n",

    " We will load the featurized data."

   ]

  },

  {

   "cell_type": "code",

   "execution_count": 38,

   "id": "c5c736f5",

   "metadata": {},

   "outputs": [],

   "source": [

    "x_data = np.load('featurized_data.npy', allow_pickle = True)\n",

    "y_data = np.load('labels.npy', allow_pickle = True)"

   ]

  },

  {

   "cell_type": "markdown",

   "id": "e31cc28d",

   "metadata": {},

   "source": [

    "## 3. Machine Learning methods and performace evaluation"

   ]

  },

  {

   "cell_type": "markdown",

   "id": "44a7a69f",

   "metadata": {},

   "source": [

    "In the modeling part, we will train classical machine learning algorithms on featurized data. For the purpose of this project, we will train the following algorithms.\n",

    "\n",

    "1. Random Forest\n",

    "2. Support Vector Classifier\n",

    "3. Logistic Regression\n",

    "4. k Nearest Neighbors"

   ]

  },

  {

   "cell_type": "markdown",

   "id": "048b9906",

   "metadata": {},

   "source": [

    "Let us train the models directly with default setting. Here we will report the 5 fold accuracy."

   ]

  },

  {

   "cell_type": "markdown",

   "id": "56fa9350",

   "metadata": {},

   "source": [

    "Random Forest"

   ]

  },

  {

   "cell_type": "code",

   "execution_count": 39,

   "id": "1798e464",

   "metadata": {},

   "outputs": [],

   "source": [

    "rf = RandomForestClassifier()\n",

    "rf_f_scores = cross_val_score(rf, x_data, y_data, cv=5)\n",

    "rf_f_acc = np.mean(rf_f_scores)"

   ]

  },

  {

   "cell_type": "markdown",

   "id": "3c628635",

   "metadata": {},

   "source": [

    "Support Vector Classifier"

   ]

  },

  {

   "cell_type": "code",

   "execution_count": 40,

   "id": "cc8b8a61",

   "metadata": {},

   "outputs": [],

   "source": [

    "svc = SVC()\n",

    "svc_f_scores = cross_val_score(svc, x_data, y_data, cv=5)\n",

    "svc_f_acc = np.mean(svc_f_scores)"

   ]

  },

  {

   "cell_type": "markdown",

   "id": "d4d78db6",

   "metadata": {},

   "source": [

    "Logistic Regression"

   ]

  },

  {

   "cell_type": "code",

   "execution_count": 41,

   "id": "710b080a",

   "metadata": {},

   "outputs": [],

   "source": [

    "lr = LogisticRegression(solver='liblinear')\n",

    "lr_f_scores = cross_val_score(lr, x_data, y_data, cv=5)\n",

    "lr_f_acc = np.mean(lr_f_scores)"

   ]

  },

  {

   "cell_type": "markdown",

   "id": "6fbb0adc",

   "metadata": {},

   "source": [

    "k Nearest Neighbors"

   ]

  },

  {

   "cell_type": "code",

   "execution_count": 42,

   "id": "48bcda5f",

   "metadata": {},

   "outputs": [],

   "source": [

    "knn = KNeighborsClassifier()\n",

    "knn_f_scores = cross_val_score(knn, x_data, y_data, cv=5)\n",

    "knn_f_acc = np.mean(knn_f_scores)"

   ]

  },

  {

   "cell_type": "markdown",

   "id": "53cc6ed6",

   "metadata": {},

   "source": [

    "#### Result on complete dataset"

   ]

  },

  {

   "cell_type": "code",

   "execution_count": 43,

   "id": "cf0345ea",

   "metadata": {},

   "outputs": [

    {

     "data": {

      "image/png": "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\n",

      "text/plain": [

       "<Figure size 720x360 with 1 Axes>"

      ]

     },

     "metadata": {},

     "output_type": "display_data"

    }

   ],

   "source": [

    "data_r = {'RF':rf_f_acc, 'SVC':svc_f_acc, 'LR':lr_f_acc, 'kNN':knn_f_acc}\n",

    "algorithm = list(data_r.keys())\n",

    "accuracy = list(data_r.values())\n",

    "fig = plt.figure(figsize = (10, 5))\n",

    "plt.bar(algorithm, accuracy, color ='red', width = 0.4)\n",

    "plt.xlabel(\"ML models\", fontsize = 18)\n",

    "plt.ylabel(\"5 fold accuracy\", fontsize = 18)\n",

    "plt.title(\"Result\", fontsize = 18)\n",

    "plt.xticks(fontsize = 14)\n",

    "plt.yticks(fontsize = 14)\n",

    "plt.ylim([0, 1])\n",

    "plt.show()"

   ]

  },

  {

   "cell_type": "markdown",

   "id": "073a25a0",

   "metadata": {},

   "source": [

    "The maximum accuracy is achieved by Random Forest. Following are the algorithm wise result."

   ]

  },

  {

   "cell_type": "code",

   "execution_count": 44,

   "id": "f4fd13c6",

   "metadata": {},

   "outputs": [

    {

     "name": "stdout",

     "output_type": "stream",

     "text": [

      "Random Forest Accuracy:  83.57142857142857\n",

      "Support Vector Classifier Accuracy:  12.142857142857144\n",

      "Logistic Regression Accuracy:  44.28571428571429\n",

      "K Nearest Neighbours Accuracy:  12.857142857142856\n"

     ]

    }

   ],

   "source": [

    "print('Random Forest Accuracy: ', rf_f_acc*100)\n",

    "print('Support Vector Classifier Accuracy: ', svc_f_acc*100)\n",

    "print('Logistic Regression Accuracy: ', lr_f_acc*100)\n",

    "print('K Nearest Neighbours Accuracy: ', knn_f_acc*100)"

   ]

  },

  {

   "cell_type": "markdown",

   "id": "a162dc00",

   "metadata": {},

   "source": [

    "We will analyze the performance of Random Forest algorithm. We will train the algorithm again with a designated test set. We want to check class-wise prediction."

   ]

  },

  {

   "cell_type": "markdown",

   "id": "dd42a75e",

   "metadata": {},

   "source": [

    "Following cell creates trainset and testset from our complete data"

   ]

  },

  {

   "cell_type": "code",

   "execution_count": 45,

   "id": "8356faea",

   "metadata": {},

   "outputs": [],

   "source": [

    "X_train = []\n",

    "X_test = []\n",

    "y_train = []\n",

    "y_test = []\n",

    "for i in range(7):\n",

    "    current_class_data = x_data[i*20: i*20 + 20]\n",

    "    X_train.append(current_class_data[0: 16])\n",

    "    X_test.append(current_class_data[16: ])\n",

    "    current_class_labels = y_data[i*20: i*20 + 20]\n",

    "    y_train.append(current_class_labels[0: 16])\n",

    "    y_test.append(current_class_labels[16: ])\n",

    "X_train = np.array(X_train).reshape(-1, 320)\n",

    "X_test = np.array(X_test).reshape(-1, 320)\n",

    "y_train = np.array(y_train).reshape(-1)\n",

    "y_test = np.array(y_test).reshape(-1)"

   ]

  },

  {

   "cell_type": "markdown",

   "id": "970f28eb",

   "metadata": {},

   "source": [

    "Lets train the model"

   ]

  },

  {

   "cell_type": "code",

   "execution_count": 46,

   "id": "41feae7a",

   "metadata": {},

   "outputs": [

    {

     "name": "stdout",

     "output_type": "stream",

     "text": [

      "Accuracy:  0.8571428571428571\n"

     ]

    }

   ],

   "source": [

    "rf = RandomForestClassifier()\n",

    "rf.fit(X_train, y_train)\n",

    "predictions = rf.predict(X_test)\n",

    "accuracy = accuracy_score(predictions, y_test)\n",

    "print('Accuracy: ', accuracy)"

   ]

  },

  {

   "cell_type": "code",

   "execution_count": 47,

   "id": "cc1970db",

   "metadata": {},

   "outputs": [

    {

     "data": {

      "image/png": "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\n",

      "text/plain": [

       "<Figure size 720x504 with 2 Axes>"

      ]

     },

     "metadata": {},

     "output_type": "display_data"

    }

   ],

   "source": [

    "conf_matrix = confusion_matrix(y_test, predictions)\n",

    "df_cm = pd.DataFrame(conf_matrix, index = [i for i in \"0123456\"], columns = [i for i in \"0123456\"])\n",

    "plt.figure(figsize = (10,7))\n",

    "sn.set(font_scale=1.4)\n",

    "sn.heatmap(df_cm, annot=True, annot_kws={\"size\": 16})\n",

    "plt.ylabel('True label')\n",

    "plt.xlabel('Predicted label')\n",

    "plt.show()"

   ]

  },

  {

   "cell_type": "markdown",

   "id": "8d59746a",

   "metadata": {},

   "source": [

    "We initially discussed the conditions under which we collected the data. Here, the indexing starts with zero (Normal condition).\n",

    "\n",

    "As we can see that the model is mainly getting confused between class 4 and class 5. After looking at the data, we realized that the class 5 fault is same as class 4 with just one bolt loosened. \n",

    "\n",

    "class 4: 1 brace on 3rd story is broken\n",

    "\n",

    "class 5: 1 brace on 3rd story is broken + unscrew the element 18\n",

    "\n",

    "The signals generated for class 5 and class 4 are very similar. As a result, the features are not distinguishable for these two classes. Hence, our model's performance is affected."

   ]

  },

  {

   "cell_type": "markdown",

   "id": "c858ebc4",

   "metadata": {},

   "source": [

    "Let us drop the class 4 datapoints and just train on 6 classes to see of our contention is correct."

   ]

  },

  {

   "cell_type": "code",

   "execution_count": 31,

   "id": "ce0cff8e",

   "metadata": {},

   "outputs": [],

   "source": [

    "idx = (y_data != 4)\n",

    "x_data = x_data[idx]\n",

    "y_data = np.array([i for i in range(6) for j in range(20)])"

   ]

  },

  {

   "cell_type": "markdown",

   "id": "ca695c49",

   "metadata": {},

   "source": [

    "Random Forest"

   ]

  },

  {

   "cell_type": "code",

   "execution_count": 32,

   "id": "24d0cb07",

   "metadata": {},

   "outputs": [],

   "source": [

    "rf = RandomForestClassifier()\n",

    "rf_f_scores = cross_val_score(rf, x_data, y_data, cv=5)\n",

    "rf_f_acc = np.mean(rf_f_scores)"

   ]

  },

  {

   "cell_type": "markdown",

   "id": "55f082bf",

   "metadata": {},

   "source": [

    "Support Vector Classifier"

   ]

  },

  {

   "cell_type": "code",

   "execution_count": 33,

   "id": "a9a5c82c",

   "metadata": {},

   "outputs": [],

   "source": [

    "svc = SVC()\n",

    "svc_f_scores = cross_val_score(svc, x_data, y_data, cv=5)\n",

    "svc_f_acc = np.mean(svc_f_scores)"

   ]

  },

  {

   "cell_type": "markdown",

   "id": "5a08c633",

   "metadata": {},

   "source": [

    "Logistic Regression"

   ]

  },

  {

   "cell_type": "code",

   "execution_count": 34,

   "id": "283c3c3d",

   "metadata": {},

   "outputs": [],

   "source": [

    "lr = LogisticRegression(solver='liblinear')\n",