13_temporal_clustering.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 59,
   "metadata": {
    "collapsed": true,
    "jupyter": {
     "outputs_hidden": true
    }
   },
   "outputs": [],
   "source": [
    "'''Main'''\n",
    "import numpy as np\n",
    "import pandas as pd\n",
    "import os, time, re\n",
    "import pickle, gzip, datetime\n",
    "from os import listdir, walk\n",
    "from os.path import isfile, join\n",
    "\n",
    "'''Data Viz'''\n",
    "import matplotlib.pyplot as plt\n",
    "import seaborn as sns\n",
    "color = sns.color_palette()\n",
    "import matplotlib as mpl\n",
    "from mpl_toolkits.axes_grid1 import Grid\n",
    "\n",
    "%matplotlib inline\n",
    "\n",
    "'''Data Prep and Model Evaluation'''\n",
    "from sklearn import preprocessing as pp\n",
    "from sklearn.model_selection import train_test_split \n",
    "from sklearn.model_selection import StratifiedKFold \n",
    "from sklearn.metrics import log_loss, accuracy_score\n",
    "from sklearn.metrics import precision_recall_curve, average_precision_score\n",
    "from sklearn.metrics import roc_curve, auc, roc_auc_score, mean_squared_error\n",
    "from keras.utils import to_categorical\n",
    "from sklearn.metrics import adjusted_rand_score\n",
    "import random\n",
    "\n",
    "'''Algos'''\n",
    "from kshape.core import kshape, zscore\n",
    "import tslearn\n",
    "from tslearn.utils import to_time_series_dataset\n",
    "from tslearn.clustering import KShape, TimeSeriesScalerMeanVariance\n",
    "from tslearn.clustering import TimeSeriesKMeans\n",
    "import hdbscan\n",
    "\n",
    "'''TensorFlow and Keras'''\n",
    "import tensorflow as tf\n",
    "import keras\n",
    "from keras import backend as K\n",
    "from keras.models import Sequential, Model\n",
    "from keras.layers import Activation, Dense, Dropout, Flatten, Conv2D, MaxPool2D\n",
    "from keras.layers import LeakyReLU, Reshape, UpSampling2D, Conv2DTranspose\n",
    "from keras.layers import BatchNormalization, Input, Lambda\n",
    "from keras.layers import Embedding, Flatten, dot\n",
    "from keras import regularizers\n",
    "from keras.losses import mse, binary_crossentropy\n",
    "from IPython.display import SVG\n",
    "from keras.utils.vis_utils import model_to_dot\n",
    "from keras.optimizers import Adam, RMSprop\n",
    "from tensorflow.examples.tutorials.mnist import input_data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "collapsed": true,
    "jupyter": {
     "outputs_hidden": true
    }
   },
   "outputs": [],
   "source": [
    "# Load the datasets\n",
    "current_path = os.getcwd()\n",
    "file = os.path.sep.join(['', 'datasets', 'ucr_time_series_data', ''])\n",
    "data_train = np.loadtxt(current_path+file+\n",
    "                        \"ECGFiveDays/ECGFiveDays_TRAIN\", \n",
    "                        delimiter=\",\")\n",
    "X_train = to_time_series_dataset(data_train[:, 1:])\n",
    "y_train = data_train[:, 0].astype(np.int)\n",
    "\n",
    "data_test = np.loadtxt(current_path+file+\n",
    "                       \"ECGFiveDays/ECGFiveDays_TEST\", \n",
    "                       delimiter=\",\")\n",
    "X_test = to_time_series_dataset(data_test[:, 1:])\n",
    "y_test = data_test[:, 0].astype(np.int)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Number of time series: 23\n",
      "Number of unique classes: 2\n",
      "Time series length: 136\n"
     ]
    }
   ],
   "source": [
    "# Basic summary statistics\n",
    "print(\"Number of time series:\", len(data_train))\n",
    "print(\"Number of unique classes:\", len(np.unique(data_train[:,0])))\n",
    "print(\"Time series length:\", len(data_train[0,1:]))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Number of time series in class 1.0: 14\n",
      "Number of time series in class 2.0: 9\n"
     ]
    }
   ],
   "source": [
    "# Number of examples in each class in the training set\n",
    "print(\"Number of time series in class 1.0:\", \n",
    "      len(data_train[data_train[:,0]==1.0]))\n",
    "print(\"Number of time series in class 2.0:\", \n",
    "      len(data_train[data_train[:,0]==2.0]))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Plot  0  Class  1.0\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Plot  1  Class  1.0\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Plot  4  Class  1.0\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Plot  7  Class  1.0\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Plot  9  Class  1.0\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Examples of Class 1.0\n",
    "for i in range(0,10):\n",
    "    if data_train[i,0]==1.0:\n",
    "        print(\"Plot \",i,\" Class \",data_train[i,0])\n",
    "        plt.plot(data_train[i])\n",
    "        plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Plot  2  Class  2.0\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Plot  3  Class  2.0\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXYAAAD8CAYAAABjAo9vAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDIuMi4zLCBodHRwOi8vbWF0cGxvdGxpYi5vcmcvIxREBQAAIABJREFUeJzt3Xl8nGW5//HPNXv2ZmmbpGmapHRfKCWFFihr2REE4YgKiIgVF9TjUVzqctSf56gcRT0Hq5XtJbIJsgiCCLIUCnSn+5auadOkSZfsk0wy9++PmaRpm2W2ZOaZXO/XKy+amcmTqw+Zb+9cz/3ctxhjUEoplTxs8S5AKaVUbGmwK6VUktFgV0qpJKPBrpRSSUaDXSmlkowGu1JKJRkNdqWUSjIa7EoplWQ02JVSKsk44vFN8/LyTElJSTy+tVJKWdbq1avrjDEjB3pdXIK9pKSEVatWxeNbK6WUZYnI3lBep60YpZRKMhrsSimVZDTYlVIqyWiwK6VUktFgV0qpJBN1sIvIWBF5U0S2iMgmEflqLApTSikVmVhMd+wA/sMYs0ZEMoDVIvKaMWZzDI6tlFIqTFGP2I0xB40xa4J/bgS2AGOiPW5v/rWlht+9VTEYh1ZKqaQR0x67iJQAZwDLe3luoYisEpFVtbW1ER3/7e21LFm6K6oalUoES7fX8u2/rufRD/ays7Yp3uWoJBOzO09FJB34K/A1Y0zDyc8bY5YASwDKy8sj2kHb47TT2t4ZVZ1KxdsbW2v4/KOrERGeXFmJTeBPd5zNeRPy4l2aShIxGbGLiJNAqD9mjHk2Fsfsjcdpp63DjzER/bugVNy9u6OOu/68hsn5maxctICl37yIouxUfvTiJnyd/niXp5JELGbFCPAgsMUY86voS+qbxxkot61D3wDKejr9hnueWUdJbiqPfvYsslKcFOem8r2rp7DjUBN//iCkZUCUGlAsRuznArcCF4vIh8GPq2Jw3FN4HHYAvD5txyjreXv7Iarqvfz7gomMSHV1P37p1NHMn5DHfa9t53BTWxwrVMkiFrNi3jXGiDFmpjFmVvDj5VgUd7IUVyDYWzXYlQU9saKSvHQXl0wZfcLjIsIPrplKg7eDJ1dWxqk6lUwsdedpVyvG69NWjLKW6novb2w9xI1njsXlOPVtN2F0BtMKM3l7e2QzxpTqyVrBrq0YZVFPr6qk02+4ec7YPl8zf8JI1u47SlNbxxBWppKRtYLdqa0YZT2dfsNTqyo5Z3wuJXlpfb7u/Al5+DoNy3cdHsLqVDKyZLDriF1ZyT82VrP/aCu3zB3X7+vOLMnG47Txzo66IapMJSuLBXtwuqP22JVFGGNY/HYFpXlpXD4tv9/Xuh125pblsnSH9tlVdCwW7DpiV9ayrOIwGw808Pnzy7DbZMDXn3daHrtqm9l/tGUIqlPJypLBrj12ZRWL365gVIab62eHti7e+RMDG9C/q+0YFQVLBXtK94hdWzEq8a2rPMayisN89rxS3MEZXQOZMCqd0Zlu3q3QYFeRs1SwH5/HriP2ZODr9LP4rZ20tCfn9L7fv72TDI+DT55dHPLXiAhzSnJYs/foIFamkp3Fgj04Yu/QYE8Gyyrq+Pk/tvL+zuSb3rertol/bKrmtnnjyPA4w/ra2cXZVNV7OVjfOkjVqWRnqWB3B+/Y8+rSvUlhe00jQFKuarhk6S5cdhu3n1Ma9tfOHpcNwJq9x2JdlhomLBXsIoLHacOrqzsmha3VgWBv70yuZZhrGrw8u+YAN5UXMTLDHfbXTy3IxO2wsWaftmNUZCwV7BBox2iPPTlsCwa7L8n+oX7o3d10+P0snD8+oq93OWzMLMpitfbZVYSsF+wODfZk0NHpZ8ehwJZwydSKqW/x8ecP9nLNzEKKc1MjPs7s4mw2VdXrz7qKiOWCPcVlp1WnO1re3iMttAdH6j5/8rRi/rx8L83tndx1QWSj9S6zx2Xj6zRsqqqPUWVqOLFcsLsdNh3FJIGuNgwkTyvG6+vk4WW7uXDSSKYWZkZ1rNnFegFVRS5We54+JCKHRGRjLI7XH+2xJ4etPYM9SVoxT6/eT11Te9SjdYCRGW7G5qRon11FJFYj9keAK2J0rH55nDZdBCwJbK9upCg7BYCOJGjFeH2dLH6zgjOKR3B2aU5MjnlmcTar9x3VzdtV2GIS7MaYpcCRWBxrIClOu64VkwS21TQyvTALoLvXbmWPvLeHqnov91w+mcD+7tE7qzSX2sY2dtc1x+R4avgYsh67iCwUkVUisqq2NvJlSbUVY32t7Z3sOdzMpPwMnHaxfCvmaHM7979ZwcWTRzFvfG7Mjju3LDDy/2DXkIyZVBIZsmA3xiwxxpQbY8pHjhwZ8XE8TrsuKWBxFYeaMAYm52fgsNksH+z3v1lBc1sH37pickyPW5qXxsgMN8t3J9+SC2pwWW5WjMdp09UdLW5rdQNAjxG7dXvIS7fX8sh7e7jxzCIm5WfE9NgiwtyyXD7YdVj77CosFgx2u64VY3EVh5pwOWyMy03D5bDuiH1zVQNffGwNp41K5/vXTB2U7zG3LIeahjb2HNaNN1ToYjXd8QngfWCSiOwXkc/G4ri90VaM9e041ERZXhp2m1i2FbPxQD13PLKSdLeDhz8zJ+wVHEM1tyzQs/9AN7hWYYjVrJhPGGMKjDFOY0yRMebBWBy3Nx6HHV+nocOCYaACKg41cdqodACcDqHDQq2Y9g4///uvHXz0/mX4jeHhz8yhICtl0L5fWbDPrsGuwuGIdwHhSnEFl+7t8JNut1wnadjz+jqpPNrCx2YXAeC022i3wD/S26obeejd3byy8SAN3g6uPb2QH183jRGprkH9vif32WM1lVIlN8sFe88NrdPdlit/2NtZG5gR0zVid9kTpxWzu66ZJ1bs487zShmV6el+fGdtEx9f8j6+Dj+XTcvnhtljmD8h8pld4ZpblsOL66rYcaiJiaNje4FWJSfLDXk9juPBrqynIriiY1ewOxJkVoyv08+XH1/DkqW7uOzXS3lxXRW+Tj+1jW3c/vAKHDbhla+ez30fnzWkoQ5w2dR87Dbh+bUHhvT7Kuuy3JDX3b3vaWKM8lR4Kg41YbcJJXmBJW2dCTJi/92bO9lU1cCiq6bw4voq7n5iLSKBgYTB8NTCeVEtwxuNkRlu5k/I44UPq/jGZZOw2bQdo/pnuWBPceqI3cp21DQxLicVd/A3r6EOdq+vkw92HWbv4RYO1ntx2YXMFCf/+8YOrptVyOfOL+P2c0t4ecNBdtU2c6jRy7Wnj+H0sSOGrMbeXH/GGL765Ics330kpne3quRkuWD3aLBbWkVtE+ODbRgAp10G9bcvYwyVR1pZve8I7+yo47VNNTS2dXR/706/wW9gVIab//zItODjNq6bNWbQaorEZVPzSXc7eHbNfg12NSALB3v8f31X4fF1+tlT18xlU0d3P+a022jydsT8e3V0+vn7hoMsWbqLTVWBO10zPQ6umJ7P1TMLmFaYRW6ai05jqGnwkuZykJ02uDNcopHisnPF9Hxe2VjNj6+bTorLHu+SVAKzYLB39dh1xG41ew830+E33RdOoWu6Y2wvnvo6/Xz8D++zZt8xxo9M44cfmcrcslwmjs7AflJ/2oZQlB2f3nm4bpg9hmdW7+eFDw9w81nF8S5HJTDLBXtXj12X7rWek2fEwOBMd/zdmztZs+8Y/33DDD5ePjZpLjbOLc1ldvEIfvHqNi6flp/Qv2Go+LLedEftsVvWjppAsI8feTzYHTFetndTVT3/+8YOPjqrkE+cVZw0oQ5gswk/vX4G9a0+/vuVLfEuRyUwywV793THJNicYajVt/h4aX0Vh5vahvx7b69p5O8bDjJmRAppPW4sc9ptMVtSoKPTzzeeXk92mov/vHZaTI6ZaKYUZHLn/FL+smo/b2ytiXc5KkFZthXTpiP2AW2uauCBd3bR4Tcca/Xx/s46fJ2Gq2cUcP+nZg9JDfsOt/D7pTt5amUlaS47/3XDjBOej+WSAk+tqmTLwQZ+f8vsQb/VP56+eskEXt1YzR2PrOLs0hy+dNFpnD9xaG+aUonNcsHe1Ypp1aV7++X1dfKlx9dQ29jGyAw3LruN288poaG1g6dWVfKlqgamFmbG7Ps1t3WwtbqR1XuPsKOmCb+Boy3tvLXtEHab8Kmzi/n3BRNP6QvHagel5rYO7nttB2eV5HD5tPyoj5fIUl0OXrz7PJ5aWcmD7+7mtodWcOGkkXzv6imcNiq8JQfW7DtKhtvBBF2qIKlYLtiddht2m+jSvQNY/NZOdtc186c7zjphNFff4uPljQf51WvbeeDT5SEd68PKYyyrqOOm8iJGZRxfQ6XySAs/eGEjq/YepbHHlMVRGW6cdhsOu7Dw/PF85twSRvdYe6WnWLVi/vjOLuqa2lhy25nDYqGsDI+TO+eXcdu8Ev70/h5+868dLPjVUs4qyeHqmQVMys+gMCuFouyUPq8zVBxq5MbF7+EP7mY1fmQ6VfWtpLkc/PT66YzLTRvav5SKGcsFO4DHobso9WdnbROL39rJdbMKT/kVPSvVycL5Zfzyte2sqzw24B2Vxhi+9/wGNh5o4Dev7+CaYGj4DfzfGzsQET56RiFF2amU5KYxe9yIE8J/ILFoxdQ1tbFk6S6umpHP7OLsqI5lNS6HjTvnl3H9GWN4fPk+/rauih/+bVP38/mZHq6eWcD8CXkUZadQlJ3a/Vvvz17ZRprLwVcXTODVTdVsPthA4QgPGw7Uc+3/LePnH5sBCGsrj7IveKfuyAw318wsYMGU0SdcK2nv8OO0y7D4R9UKYhLsInIF8BvADjxgjPlZLI7blxSXbmjdl7qmNr78+Fo8Thvfu7r3XX0+c14pDy3bzY9e3MSTC+fhcvR9Df3DymNsPNDAFy8cT4PXxwsfVvFscDGqeWW53HvTzKjmgbuCrZholqRdur2WlvZOvnjhaRHXYXW56W7uvmQCd18ygX2HW9h3pIXKoy28sfUQj76/lwff3Q1AhtvBvTedTk6ai9e31PDNyydx5/wy7pxf1n2sfYdbWPjoKu768xogMCW1KCeFgiwPG/bX89rmGjxOG5dMGc2ccdks3VHHOztqEYT8LA8TR2dw5rhsLpw0kikFsWv3DQVjDA8t20PVsdZB2xVrKEQd7CJiB+4HLgX2AytF5G/GmM3RHrsvbodd57H3ovJIC7c+uJzqBi+/v+VMRma4e31dutvBj66bzleeWMtP/76ZH103vc9jPvrBXtJcdr5w4XgyPE7+30dn0OD1cbS5nbHZqVFPJ3TYbRgDnX6Dwx7ZsfbUNWMTdEnboOLc1O4Fyz5xVjH1rT62VTdysL6Vh5bt4a4/ryYv3U1+poc7zi3t9euf/eI5vLqpmuKcVKYVZnWP8v1+w6q9R3lxXRUvbzjI39cfpDDLw61zS3DahQPHWtlc1cDrW2r4+T+2cv0ZY/jihePZH3x8ZlEW88pyg//fAy24RBnl+/2GH7+0mUfe2wPAzXPGWvbaQyxG7GcBFcaYXQAi8iRwHTBowe5x2mgb5q2YdZXHeHVTNVXHWqk65qWqvpXqei9pbgeP3TmXM8f135K49vRC1lce44F3dzN9TBY3lY895TVHmtt5af1BPl4+9oSt3zI9TjJjtBWcM7hZSoff4IjwLvndh1soyk7t9zeP4SwrxclZpTkAXD4tn+8/v5GnV+/n3htn9rk0QarLwfVnFJ3yuM0mnFWaw1mlOfzwI1OpPNrKuJxT/4Gva2rj4WW7+eM7u3nupOWGc9NcZKe5OHislRSXg6tm5HPVjAJmjR3R/Q9IPCx6fiNPrNjHzXPG8szq/Ty9ej/fvWpK3OqJRiyCfQxQ2ePz/cDZMThunzzO4dWKqW/xcfMfP8Dr62R2cTb7j7awfPcRHLbAr76FI1IoH5dN4YgUPnZm0Qk3APXn21dOZlNVA998Zj0vfFjFXReM57wJed3P/2VVJe0dfm6dN26w/mo4g6P09k5/xG/q3XVNlOTphb5QeJx2fnHjTL5yyQTG5kS3lILDbqO0j/Oel+7mm5dP5uY5xby+pYZJozOYXJDJit2HeXlDNW0dncyfkEdNg5enVlbyp/f34rLbmFKYScZJG+iU5KXymXNLQ/65jsR7O+t4YsU+Pje/lO9eNYUjze08u+YA37x8Uvfgw0piEey9/R51yjQHEVkILAQoLo5unYuUJNrQutNvuO+17TjtNm6dN46ck6YD+v2Grz21lopDjZw/YSRvbTuE22Hje1dP4eNzxka1ibLDbuOBT5fzp/f38tCy3dzy4HJumVvMD66Zxjs7arn/jQrmluUMaouj603ji/CGM2MMe+paOHOYXTSNhohEHeqhGpsTCOUuV0wv4IrpBSe8pqmtg2UVdazee5QN++tPaLMaY3h61X4eW76Pc8fnMX5kGkXZqVw0eWT31M5DjV6cNlvESyx0dPr58YubKcpO4T8um4SIcFP5WP65uYa3t9WyoMeidVYRi2DfD/T8Pb4IqDr5RcaYJcASgPLy8qjmt3mcdlraY78i4FDz+w2LntvAkysDv/AsfruCc8fn4bALHqedGWOy2H+0lTe31fKTj07n1rnjYr7vZZrbwRcuHM8d55Xwq9e284e3d/HezsPsrmtmakEmv/y3WTH7Xr3pDvYIpzzWNbXT1NahI3YLS3c7uHxafp/3H9Q1tfGn9/fy6sZq1lUeo7Gtg5++vIWJo9Np9XVSeaQVu004Z3wul04dzdicVMZmp1Cal37Kom+9eWLFPrZWN7L4U7O7f2u8cNJI8tJd/GVV5bAN9pXABBEpBQ4ANwOfjMFx++Rx2jjSbM0e+7GWdv65uYaOTsOqPUd4du0B7r74NK49vZAH3tnNuv3HAGhoDcxAAbjpzCJuOTvwW85gXWhyO+x858opTCvM4tt/Xc/1s8bwXzfMGPSeZ1crJtKblPYcbgbosyWgrC8v3c3XL53I1y+dCEBNg5dXNhzk1U01ZKU4uW1uCcda23lx3UF+8MLxqZ7pbgezxo7oHsmPznBTXpKNy2HjpfUHea/iMB1+P/WtPuaV5XLF9OP/sDjtNq4/YwwPL9vDvsMtcds9K1JRB7sxpkNEvgy8SmC640PGmE0DfFlU3BZsxfg6/Ty+fB/3vb6dYy2+7sc/f0EZX790IiLCz2+cecLXVNd72VbTyLyy3CGbOXDt6YVcOT1/yPqKXRc8Iw323XUa7MPN6EwPt59byu0nzej5xmWTOFjv5WB9K3sPt7Bm31E+rDxG1bFWDPDqsVYeCE77zPQ4uGjyKNLdju57AU5+j91+bilPrqzkK0+u5em75lmq1x6TeezGmJeBl2NxrFCkOO14E3hJAb/fsGxnHUea2zEGVu45wisbqznS3M65p+Vyz+WTyc/y4LTbTump95Sf5SE/K/SbfWJlKH+AHbboWjG765px2IQxI1JiWZayIBGhcEQKhSNSOHNcDjfMPnFWT3uHn41V9TR5O5hbljvgLKoxI1L42Q0z+dLja7jvte3cc8XkwSw/pqx556nTFpfVHTv9pt+enTGGNfuO8ZOXNvNh5bHux1Ocdi6ZMoobzyzigokjE2bebiKIuhVT10xxTioOC42mVHy4HLaw70y+emYB7+wYy+K3dzKnNIeLJo0apOpiy5rB7hja6Y7GGH731k5+/fr2wGigOJucNBeGQC+8qr6Vg8G55F6fn5EZbn5x48zuueQFWR5SXZY81YPOGYNWjF44VYPphx+Zxrr99Xz5sTX85a55TCvMindJA7Jk2nQtKRDJDBGvr5Pf/msHQPevVtX1Xn7y0mbuumA8M4oC/9NqG9vYWh3YK/OJFft4eUM1C6aMwibCOxV1tAQ3RE73OCgckcKUgkwumTKK4tw0rj9jDOluS57aIeeMohXj9xv2Hm7h3NPyBn6xUhFKcdl5+PY5XP+7ZdzxyEqe++K5FCZ468+S6eNx2vGbwE0t7uDtitX1XlJcdrJS+p7XvWF/PV//y4fsCG7RNrUwk6tnFPCtv67n7e21vFtRxxOfm0ttUxtfeWIt9a2Bi5w2gUVXTeHO+aXaRomxaFoxNY1eWn2dOmJXgy4/y8PDn5nDTYvf52OL3+P3t5zZ5wJ6fr+horaJ1XuP0tDqo2BEChkeBzX1XqqOtXJT+dhBv4/AksHudnRtaB0I9rX7jnLrgyuwCXxtwURunTfulAuAT67Yx/df2EhOmouHb5/Dfa9v5/vPb2Tv4Rbe3l7LFy4cz/NrD/DxJe/T3NbBxNEZ3P/J2XicNkZleCw33ckqomnFdM+I0eVl1RCYnJ/Jk5+fy8I/reamP7zPvTfO5LpZY054TYPXx02L32dbTWOvxxCBWcUjNNh74+mxi9LGIy3c9tAKctNdFOek8uOXNvPfr2zBJoGbfGYWZZHmcvCPTdXMn5DHb28+g+w0F0XZKVz923e599VtzC3L4ZuXTeLfysdy64PLuWjSKH72sRnaFx8CrihuUNpT1wIEbjlXaihMK8zixbvP465HV3PPM+uZW5bbvdeAMYZFz22koraJn1w3jfMmBG5yOljvpdHrY3Smh9GZniGZdWbJ5OraHu/mP37AgaOt5KW7efxzcynM8vDW9lo+2HUYgIbWDtbuO8ryuiN8/oIy7rl8cveslgmjM/juVZNZ/PZO7r3xdGw2oTQvjXfuuUjbLUPIEUUrZu/hZlx2G4VZid3vVMklJ83F/9x0Ohf/8i1+92ZF9+qoT6/ez4vrqvjm5ZO4dV5J9+ujWfYjUpYM9lnFI5hTkk2mx8n80/K4c35Z9zzmiyaNOmVKkt9vel1e9vZzS7ltXskJz2moD63jSwqEH+wH670UjPBEvXSwUuEqzk3lpvKxPLGikoUXjGfnoSZ++MIm5pXlctcF4+NdnjWDffzIdJ6+65yQX9/fG19DIb6iacVU13v73HJPqcF298Wn8dfV+/nsIyvZXtPIxNEZ/PrmWSGtTzPY9K4OFVfRtGIONrRSEIc7c5UCKByRwifPLmZrdSNXzijg2S+ekzADDUuO2FXyiLQVY4yhpr4tLksuKNXl21dO5tKpozln/NCt5xQKDXYVV5Eu23ukuZ32Tj/5CTJCUsOTx2lPyBvktBWj4soV4Yi9usELoK0YpXqhwa7iqrvHHuaibtX1gWBPlJ6mUolEg13FlSM4g8DnD68Vc7C+a8Suc9iVOpkGu4orEcFlt4Xdiqlp8GITyEuPbJ9LpZJZVMEuIjeJyCYR8YtIeayKUsOLwy5ht2IO1nsZleHRddiV6kW074qNwA3A0hjUooYpZwQj9up6r051VKoPUU13NMZsAb0NX0XHabeF3WOvbvBy2sj0QapIKWvT32NV3LkiaMXoiF2pvg04YheR14H8Xp5aZIx5IdRvJCILgYUAxcXFIReokp8jzFZMo9dHU1uHBrtSfRgw2I0xC2LxjYwxS4AlAOXl5ZFtSa+SktMuYbViavTmJKX6pa0YFXdOuy2sVkzXHHZdTkCp3kU73fF6EdkPzAP+LiKvxqYsNZyEOyum665TbcUo1btoZ8U8BzwXo1rUMOW0S1iLgOlyAkr1T1sxKu7CHrE3eMlOdXbvfauUOpEGu4o7lyP8Vky+rhGjVJ802FXcOWzhtWJqGr2MznQPYkVKWZsGu4q7cFsxR5rayUnTxb+U6osGu4o7Z5itmCMt7eRqsCvVJw12FXfOMFoxLe0deH1+ctK0FaNUXzTYVdyF04o53NQOQE6aczBLUsrSNNhV3AVaMaGN2I80dwW7jtiV6osGu4q7cHZQOtLSFezaY1eqLxrsKu4C0x1DDPYmDXalBqLBruLO6bDREXYrRoNdqb5osKu4c9pttHf6MWbgcD/S0o7TLmR6olrmSKmkpsGu4s5pC2yt2BHCmuxHmtrJTnXpdoxK9UODXcWd0xH4MQylz364We86VWogGuwq7pz2rmAfeMR+tEWDXamBaLCruHPZA22VUEbsR3TErtSAot1B6V4R2Soi60XkOREZEavC1PDhsIfRimlq02BXagDRjthfA6YbY2YC24HvRF+SGm66WzEd/bdifJ1+GrwdGuxKDSCqYDfG/NMY0xH89AOgKPqS1HDj7GrF+PsfsR8N3nWqKzsq1b9Y9tjvAF6J4fHUMOEMsRWj68QoFZoB7/IQkdeB/F6eWmSMeSH4mkVAB/BYP8dZCCwEKC4ujqhYlZxCbcV0BXu2ruyoVL8GDHZjzIL+nheRTwPXAJeYfm4dNMYsAZYAlJeXh74Pmkp6obZiuoI9V0fsSvUrqvuyReQK4FvABcaYltiUpIYbV/eIPdRWjPbYlepPtD32/wMygNdE5EMR+X0MalLDjCPEG5S6gn1EqrZilOpPVCN2Y8xpsSpEDV/OEG9QOtLcTlaKs7snr5Tqnb5DVNyFOivmcLNuYq1UKDTYVdyFulbMUV1OQKmQaLCruAunFZOtwa7UgDTYVdxpK0ap2NJgV3HncgzcijHGaCtGqRBpsKu4c9gGbsU0t3fS4Tc61VGpEGiwq7gLZQelRq8PgAyPBrtSA9FgV3HnCmFWTKM3sIhohm5irdSANNhV3IXSitERu1Kh02BXcWe3CSL9B3uDjtiVCpkGu4o7EcFpt9HezyJgXa2YTA12pQakwa4Sgttho63fYNdWjFKh0mBXCcHtsA8Q7NqKUSpUGuwqIQRG7J19Pt/o9WG3CSlO+xBWpZQ1abCrhOBxDtSK6SDD40BEhrAqpaxJg10lBLfDTptv4GBXSg0sqmAXkZ+IyPrg7kn/FJHCWBWmhhe3c+BWTIZbL5wqFYpoR+z3GmNmGmNmAS8BP4hBTWoYGmhWTIOO2JUKWVTBboxp6PFpGtD/TglK9SGUWTE61VGp0EQ9BBKRnwK3AfXARVFXpIYlt8NGm6//VkymJ2MIK1LKugYcsYvI6yKysZeP6wCMMYuMMWOBx4Av93OchSKySkRW1dbWxu5voJKC22kf8M5TbcUoFZoB3ynGmAUhHutx4O/AD/s4zhJgCUB5ebm2bNQJ+uuxG2NoatNWjFKhinZWzIQen14LbI2uHDVc9XeDUkt7J51+oyN2pUIU7TvlZyIyCfADe4G7oi9JDUf9zWM/vpyAjth254A/AAALl0lEQVSVCkVUwW6M+VisClHDm9tpw9vHiP34AmA6YlcqFHrnqUoIbocNX6eh03/q5Rddi12p8Giwq4TgdgQW9+ptZowu2atUeDTYVUJwBze07u0Cqm6yoVR4NNhVQvAEl+PtbcqjXjxVKjwa7CohdI/Ye5kZoxdPlQqPBrtKCG5n/60Yu01IdekmG0qFQoNdJYSui6e9t2J8pLt1kw2lQqXBrhLCQBdPtQ2jVOg02FVC6K/H3qBL9ioVFg12lRDc/c6K8emIXakwaLCrhDBQK0bnsCsVOg12lRCOB3svI/Y2n7ZilAqDBrtKCN2tmF7nsevFU6XCocGuEkJfrRhjjAa7UmHSYFcJoa9WTKuva5MNbcUoFSoNdpUQum5Q8p60oXWjLtmrVNhiEuwi8g0RMSKSF4vjqeHHaRdETh2x65K9SoUv6mAXkbHApcC+6MtRw5WI9LqhtW6yoVT4YjFivw+4Bzh16xulwuBx2mk7qRXT1BXsbg12pUIVVbCLyLXAAWPMuhjVo4ax3kbsLe2BoE91abArFaoB3y0i8jqQ38tTi4DvApeF8o1EZCGwEKC4uDiMEtVw4XbYewn2wIhdl+xVKnQDBrsxZkFvj4vIDKAUWBdcTrUIWCMiZxljqns5zhJgCUB5ebm2bdQpAiP2E1sx3SN2twa7UqGK+PdbY8wGYFTX5yKyByg3xtTFoC41DLmdtlPuPD0+YtdWjFKh0nnsKmH03ooJjNhTnDpiVypUMRsGGWNKYnUsNTz11YpJcdqx23T3JKVCpSN2lTB6nxXToRdOlQqTBrtKGG6H/dQee1unXjhVKkwa7CphuJ2ntmKa2ztIdeqFU6XCocGuEkZfNyjpiF2p8Giwq4TR16wY7bErFR4NdpUw3A7bKWvFBIJdWzFKhUODXSUMt9OGV2fFKBU1DXaVMDwOO51+Q0fn8XDXEbtS4dNgVwnD7Tx1e7yWNh2xKxUuDXaVMLq2x+sKdmMMLb5O0jTYlQqLBrtKGMc3tA5cQPX6/BgDKdqKUSosGuwqYXS3YoJ3n3at7Jim89iVCosGu0oYJ7didGVHpSKjwa4SxsmtmK5gT9P9TpUKiwa7Shgnj9ibg62YFL14qlRYNNhVwjilx94WHLHrxVOlwhJVsIvIf4rIARH5MPhxVawKU8PPqa0Y3chaqUjEYih0nzHmf2JwHDXM9XXxVINdqfBoK0YljL4unuqSAkqFJxbB/mURWS8iD4lIdgyOp4apvuax63rsSoVnwGAXkddFZGMvH9cBi4HxwCzgIPDLfo6zUERWiciq2tramP0FVPLosxWj89iVCsuAv+MaYxaEciAR+SPwUj/HWQIsASgvLzehFqiGD09wxO4Nrsne3N6By2HDYdeOoVLhiHZWTEGPT68HNkZXjhrOXPYTV3dsbdcFwJSKRLRXpX4hIrMAA+wBPh91RWrYcthtOGzSffG0uU3XYlcqElG9a4wxt8aqEKWga3u84Ijdp2uxKxUJbV6qhOJ2Ht/QOjBi12BXKlwa7CqhuB227lZMq26Lp1RENNhVQgkE+/FFwHTErlT4NNhVQnE77Md77O2dpOqSvUqFTYNdJRS383grprm9Q29OUioCGuwqofRsxbS0depyAkpFQINdJZR0t4OjLT6MMbT4dFaMUpHQYFcJZVJ+JhWHGmlq66DTb3RWjFIR0GBXCWVaYSa+TsO6ynpA12JXKhIa7CqhTB+TBcCKPUcA3RZPqUhosKuEMi4nlXS3g5W7A8GuG1krFT4NdpVQbDZhakEmayuPApCms2KUCpsGu0o408Zk4g3epJTi1FaMUuHSYFcJZ3phVvefdcSuVPg02FXC6bqACrqRtVKR0GBXCWf8yDTcjsCPpk53VCp8UQe7iNwtIttEZJOI/CIWRanhzWG3MbkgE9DpjkpFIqp3jYhcBFwHzDTGtInIqNiUpYa76YWZrKs8ptMdlYpAtMOhLwA/M8a0ARhjDkVfklJw27wSxmSn4HJot1CpcEUb7BOB+SLyU8ALfMMYszL6stRwNyk/g0n5GfEuQylLGjDYReR1IL+XpxYFvz4bmAvMAf4iImXGGNPLcRYCCwGKi4ujqVkppVQ/Bgx2Y8yCvp4TkS8AzwaDfIWI+IE8oLaX4ywBlgCUl5efEvxKKaViI9oG5vPAxQAiMhFwAXXRFqWUUipy0fbYHwIeEpGNQDvw6d7aMEoppYZOVMFujGkHbolRLUoppWJA55IppVSS0WBXSqkko8GulFJJRuJxrVNEaoG9EX55HtabeaM1Dz6r1Qta81CxWs391TvOGDNyoAPEJdijISKrjDHl8a4jHFrz4LNavaA1DxWr1RyLerUVo5RSSUaDXSmlkowVg31JvAuIgNY8+KxWL2jNQ8VqNUddr+V67EoppfpnxRG7Ukqpflgq2EXkiuA2fBUi8u1413MyERkrIm+KyJbgVoFfDT6eIyKviciO4H+z413ryUTELiJrReSl4OelIrI8WPNTIuKKd409icgIEXlGRLYGz/e8RD/PIvLvwZ+LjSLyhIh4Eu08i8hDInIouP5T12O9nlcJ+G3w/bheRGYnSL33Bn8u1ovIcyIyosdz3wnWu01ELh/qevuqucdz3xARIyJ5wc8jOseWCXYRsQP3A1cCU4FPiMjU+FZ1ig7gP4wxUwisUf+lYI3fBv5ljJkA/Cv4eaL5KrClx+c/B+4L1nwU+Gxcqurbb4B/GGMmA6cTqD1hz7OIjAG+ApQbY6YDduBmEu88PwJccdJjfZ3XK4EJwY+FwOIhqrGnRzi13teA6caYmcB24DsAwffizcC04Nf8LpgrQ+0RTq0ZERkLXArs6/FwZOfYGGOJD2Ae8GqPz78DfCfedQ1Q8wvB/1HbgILgYwXAtnjXdlKdRQTesBcDLwFC4AYJR2/nPt4fQCawm+A1oh6PJ+x5BsYAlUAOgcX3XgIuT8TzDJQAGwc6r8AfgE/09rp41nvSc9cDjwX/fEJmAK8C8xLhHAcfe4bAIGUPkBfNObbMiJ3jb4wu+4OPJSQRKQHOAJYDo40xBwGC/020Tb9/DdwD+IOf5wLHjDEdwc8T7VyXEdjM5eFg++gBEUkjgc+zMeYA8D8ERmMHgXpgNYl9nrv0dV6t8J68A3gl+OeErVdErgUOGGPWnfRURDVbKdill8cSckqPiKQDfwW+ZoxpiHc9/RGRa4BDxpjVPR/u5aWJdK4dwGxgsTHmDKCZBGq79CbYl74OKAUKgTQCv2afLJHO80AS+udERBYRaI8+1vVQLy+Le70ikkpgq9Ef9PZ0L48NWLOVgn0/MLbH50VAVZxq6ZOIOAmE+mPGmGeDD9eISEHw+QLgULzq68W5wLUisgd4kkA75tfACBHpWq8/0c71fmC/MWZ58PNnCAR9Ip/nBcBuY0ytMcYHPAucQ2Kf5y59ndeEfU+KyKeBa4BPmWAPg8StdzyBf/DXBd+HRcAaEcknwpqtFOwrgQnBWQQuAhdB/hbnmk4gIgI8CGwxxvyqx1N/Az4d/POnCfTeE4Ix5jvGmCJjTAmBc/qGMeZTwJvAjcGXJVrN1UCliEwKPnQJsJkEPs8EWjBzRSQ1+HPSVXPCnuce+jqvfwNuC87cmAvUd7Vs4klErgC+BVxrjGnp8dTfgJtFxC0ipQQuSK6IR409GWM2GGNGGWNKgu/D/cDs4M95ZOc4HhcOorjgcBWBq9w7gUXxrqeX+s4j8GvSeuDD4MdVBHrW/wJ2BP+bE+9a+6j/QuCl4J/LCPzQVwBPA+5413dSrbOAVcFz/TyQnejnGfgRsBXYCDwKuBPtPANPELgG4AsGzGf7Oq8E2gT3B9+PGwjM+EmEeisI9KW73oO/7/H6RcF6twFXJso5Pun5PRy/eBrROdY7T5VSKslYqRWjlFIqBBrsSimVZDTYlVIqyWiwK6VUktFgV0qpJKPBrpRSSUaDXSmlkowGu1JKJZn/D4H8bWb+dulTAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Plot  5  Class  2.0\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Plot  6  Class  2.0\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Plot  8  Class  2.0\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Examples of Class 2.0\n",
    "for i in range(0,10):\n",
    "    if data_train[i,0]==2.0:\n",
    "        print(\"Plot \",i,\" Class \",data_train[i,0])\n",
    "        plt.plot(data_train[i])\n",
    "        plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {
    "collapsed": true,
    "jupyter": {
     "outputs_hidden": true
    }
   },
   "outputs": [],
   "source": [
    "# Prepare the data - Scale\n",
    "X_train = TimeSeriesScalerMeanVariance(mu=0., std=1.).fit_transform(X_train)\n",
    "X_test = TimeSeriesScalerMeanVariance(mu=0., std=1.).fit_transform(X_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true,
    "jupyter": {
     "outputs_hidden": true
    }
   },
   "outputs": [],
   "source": [
    "# k-Shape Algorithm"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "KShape(max_iter=100, n_clusters=2, n_init=100, random_state=None, tol=1e-06,\n",
       "    verbose=0)"
      ]
     },
     "execution_count": 42,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Train using k-Shape\n",
    "ks = KShape(n_clusters=2, max_iter=100, n_init=100,verbose=0)\n",
    "ks.fit(X_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Adjusted Rand Index: 0.668041237113402\n"
     ]
    }
   ],
   "source": [
    "# Make predictions on train set and calculate adjusted Rand index\n",
    "preds = ks.predict(X_train)\n",
    "ars = adjusted_rand_score(data_train[:,0],preds)\n",
    "print(\"Adjusted Rand Index:\", ars)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Adjusted Rand Index on Test Set: 0.0006332050676187496\n"
     ]
    }
   ],
   "source": [
    "# Make predictions on test set and calculate adjusted Rand index\n",
    "preds_test = ks.predict(X_test)\n",
    "ars = adjusted_rand_score(data_test[:,0],preds_test)\n",
    "print(\"Adjusted Rand Index on Test Set:\", ars)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {
    "collapsed": true,
    "jupyter": {
     "outputs_hidden": true
    }
   },
   "outputs": [],
   "source": [
    "# ECG 5000"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "metadata": {
    "collapsed": true,
    "jupyter": {
     "outputs_hidden": true
    }
   },
   "outputs": [],
   "source": [
    "# Load the datasets\n",
    "current_path = os.getcwd()\n",
    "file = os.path.sep.join(['', 'datasets', 'ucr_time_series_data', ''])\n",
    "data_train = np.loadtxt(current_path+file+\n",
    "                        \"ECG5000/ECG5000_TRAIN\", \n",
    "                        delimiter=\",\")\n",
    "\n",
    "data_test = np.loadtxt(current_path+file+\n",
    "                       \"ECG5000/ECG5000_TEST\", \n",
    "                       delimiter=\",\")\n",
    "\n",
    "data_joined = np.concatenate((data_train,data_test),axis=0)\n",
    "data_train, data_test = train_test_split(data_joined, \n",
    "                                    test_size=0.20, random_state=2019)\n",
    "\n",
    "X_train = to_time_series_dataset(data_train[:, 1:])\n",
    "y_train = data_train[:, 0].astype(np.int)\n",
    "X_test = to_time_series_dataset(data_test[:, 1:])\n",
    "y_test = data_test[:, 0].astype(np.int)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Number of time series: 4000\n",
      "Number of unique classes: 5\n",
      "Time series length: 140\n"
     ]
    }
   ],
   "source": [
    "# Summary statistics\n",
    "print(\"Number of time series:\", len(data_train))\n",
    "print(\"Number of unique classes:\", len(np.unique(data_train[:,0])))\n",
    "print(\"Time series length:\", len(data_train[0,1:]))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Number of time series in class 1.0: 2327\n",
      "Number of time series in class 2.0: 1423\n",
      "Number of time series in class 3.0: 75\n",
      "Number of time series in class 4.0: 156\n",
      "Number of time series in class 5.0: 19\n"
     ]
    }
   ],
   "source": [
    "# Calculate number of readings per class\n",
    "print(\"Number of time series in class 1.0:\", \n",
    "      len(data_train[data_train[:,0]==1.0]))\n",
    "print(\"Number of time series in class 2.0:\", \n",
    "      len(data_train[data_train[:,0]==2.0]))\n",
    "print(\"Number of time series in class 3.0:\", \n",
    "      len(data_train[data_train[:,0]==3.0]))\n",
    "print(\"Number of time series in class 4.0:\", \n",
    "      len(data_train[data_train[:,0]==4.0]))\n",
    "print(\"Number of time series in class 5.0:\", \n",
    "      len(data_train[data_train[:,0]==5.0]))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      " Class  1.0  Count  2327\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      " Class  2.0  Count  1423\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      " Class  3.0  Count  75\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      " Class  4.0  Count  156\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      " Class  5.0  Count  19\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Display readings from each class\n",
    "for j in np.unique(data_train[:,0]):\n",
    "    dataPlot = data_train[data_train[:,0]==j]\n",
    "    cnt = len(dataPlot)\n",
    "    dataPlot = dataPlot[:,1:].mean(axis=0)\n",
    "    print(\" Class \",j,\" Count \",cnt)\n",
    "    plt.plot(dataPlot)\n",
    "    plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "metadata": {
    "collapsed": true,
    "jupyter": {
     "outputs_hidden": true
    }
   },
   "outputs": [],
   "source": [
    "# Prepare data - Scale\n",
    "X_train = TimeSeriesScalerMeanVariance(mu=0., std=1.).fit_transform(X_train)\n",
    "X_test = TimeSeriesScalerMeanVariance(mu=0., std=1.).fit_transform(X_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 50,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Init 1\n",
      "0.024 --> 0.021 --> 0.020 --> 0.019 --> 0.019 --> 0.019 --> 0.019 --> \n",
      "Init 2\n",
      "0.041 --> 0.032 --> 0.030 --> 0.029 --> 0.029 --> 0.028 --> 0.028 --> 0.028 --> 0.027 --> 0.022 --> 0.018 --> 0.018 --> 0.017 --> 0.017 --> 0.018 --> \n",
      "Init 3\n",
      "0.024 --> 0.017 --> 0.017 --> \n",
      "Init 4\n",
      "0.028 --> 0.019 --> 0.018 --> 0.018 --> 0.018 --> 0.018 --> \n",
      "Init 5\n",
      "0.049 --> 0.032 --> 0.032 --> 0.031 --> 0.031 --> 0.030 --> 0.030 --> 0.030 --> \n",
      "Init 6\n",
      "0.044 --> 0.023 --> 0.019 --> 0.017 --> 0.017 --> 0.017 --> \n",
      "Init 7\n",
      "0.039 --> 0.027 --> 0.020 --> 0.019 --> 0.019 --> 0.018 --> 0.018 --> 0.018 --> \n",
      "Init 8\n",
      "0.032 --> 0.018 --> 0.018 --> 0.018 --> \n",
      "Init 9\n",
      "0.023 --> 0.019 --> 0.018 --> 0.019 --> \n",
      "Init 10\n",
      "0.064 --> 0.057 --> 0.055 --> 0.054 --> 0.051 --> 0.049 --> 0.046 --> 0.043 --> 0.041 --> 0.038 --> 0.036 --> 0.034 --> 0.033 --> 0.032 --> 0.031 --> 0.031 --> 0.031 --> 0.031 --> \n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "KShape(max_iter=100, n_clusters=5, n_init=10, random_state=2019, tol=1e-06,\n",
       "    verbose=1)"
      ]
     },
     "execution_count": 50,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Train using k-Shape\n",
    "ks = KShape(n_clusters=5, max_iter=100, n_init=10,verbose=1,random_state=2019)\n",
    "ks.fit(X_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 51,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Adjusted Rand Index on Training Set: 0.7499312374127193\n"
     ]
    }
   ],
   "source": [
    "# Predict on train set and calculate adjusted Rand index\n",
    "preds = ks.predict(X_train)\n",
    "ars = adjusted_rand_score(data_train[:,0],preds)\n",
    "print(\"Adjusted Rand Index on Training Set:\", ars)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Adjusted Rand Index on Test Set: 0.7172302400677499\n"
     ]
    }
   ],
   "source": [
    "# Predict on test set and calculate adjusted Rand index\n",
    "preds_test = ks.predict(X_test)\n",
    "ars = adjusted_rand_score(data_test[:,0],preds_test)\n",
    "print(\"Adjusted Rand Index on Test Set:\", ars)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 53,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Predicted Cluster  0.0\n",
      "2.0    29\n",
      "4.0     2\n",
      "1.0     2\n",
      "3.0     2\n",
      "5.0     1\n",
      "Name: actual, dtype: int64\n",
      "\n",
      "Predicted Cluster  1.0\n",
      "2.0    270\n",
      "4.0     14\n",
      "3.0      8\n",
      "1.0      2\n",
      "5.0      1\n",
      "Name: actual, dtype: int64\n",
      "\n",
      "Predicted Cluster  2.0\n",
      "1.0    553\n",
      "4.0     16\n",
      "2.0      9\n",
      "3.0      7\n",
      "Name: actual, dtype: int64\n",
      "\n",
      "Predicted Cluster  3.0\n",
      "2.0    35\n",
      "1.0     5\n",
      "4.0     5\n",
      "5.0     3\n",
      "3.0     3\n",
      "Name: actual, dtype: int64\n",
      "\n",
      "Predicted Cluster  4.0\n",
      "1.0    30\n",
      "4.0     1\n",
      "3.0     1\n",
      "2.0     1\n",
      "Name: actual, dtype: int64\n",
      "\n",
      "Count of Non-Primary Points:  83\n"
     ]
    }
   ],
   "source": [
    "# Evaluate goodness of the clusters\n",
    "preds_test = preds_test.reshape(1000,1)\n",
    "preds_test = np.hstack((preds_test,data_test[:,0].reshape(1000,1)))\n",
    "preds_test = pd.DataFrame(data=preds_test)\n",
    "preds_test = preds_test.rename(columns={0: 'prediction', 1: 'actual'})\n",
    "\n",
    "counter = 0\n",
    "for i in np.sort(preds_test.prediction.unique()):\n",
    "    print(\"Predicted Cluster \", i)\n",
    "    print(preds_test.actual[preds_test.prediction==i].value_counts())\n",
    "    print()\n",
    "    cnt = preds_test.actual[preds_test.prediction==i] \\\n",
    "                        .value_counts().iloc[1:].sum()\n",
    "    counter = counter + cnt\n",
    "print(\"Count of Non-Primary Points: \", counter)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 455,
   "metadata": {
    "collapsed": true,
    "jupyter": {
     "outputs_hidden": true
    }
   },
   "outputs": [],
   "source": [
    "# Time Series k-Means"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 54,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Init 1\n",
      "36.241 --> 27.298 --> 26.273 --> 25.556 --> 25.120 --> 24.911 --> 24.831 --> 24.815 --> 24.813 --> 24.812 --> 24.812 --> 24.812 --> 24.812 --> \n",
      "Init 3\n",
      "37.226 --> 25.264 --> 24.680 --> 24.430 --> 24.296 --> 24.202 --> 24.155 --> 24.136 --> 24.128 --> 24.124 --> 24.121 --> 24.119 --> 24.118 --> 24.117 --> 24.117 --> 24.117 --> 24.117 --> 24.116 --> 24.116 --> 24.116 --> 24.116 --> 24.115 --> 24.115 --> 24.115 --> 24.114 --> 24.113 --> 24.112 --> 24.111 --> 24.110 --> 24.109 --> 24.108 --> 24.108 --> 24.108 --> \n",
      "Init 5\n",
      "33.738 --> 25.341 --> 25.014 --> 24.795 --> 24.664 --> 24.623 --> 24.616 --> 24.615 --> 24.614 --> 24.614 --> 24.614 --> \n",
      "Init 7\n",
      "30.531 --> 24.956 --> 24.740 --> 24.663 --> 24.629 --> 24.615 --> 24.603 --> 24.595 --> 24.585 --> 24.567 --> 24.560 --> 24.559 --> 24.559 --> 24.559 --> 24.559 --> 24.559 --> 24.559 --> \n",
      "Init 9\n",
      "34.030 --> 25.598 --> 25.398 --> 25.344 --> 25.310 --> 25.277 --> 25.236 --> 25.159 --> 25.005 --> 24.834 --> 24.725 --> 24.666 --> 24.636 --> 24.620 --> 24.609 --> 24.600 --> 24.594 --> 24.588 --> 24.583 --> 24.579 --> 24.575 --> 24.572 --> 24.571 --> 24.569 --> 24.568 --> 24.567 --> 24.566 --> 24.566 --> 24.565 --> 24.565 --> 24.565 --> 24.565 --> 24.565 --> 24.565 --> \n",
      "Init 11\n",
      "35.921 --> 26.806 --> 26.175 --> 25.948 --> 25.847 --> 25.790 --> 25.758 --> 25.737 --> 25.728 --> 25.724 --> 25.720 --> 25.717 --> 25.710 --> 25.706 --> 25.706 --> 25.704 --> 25.704 --> 25.703 --> 25.702 --> 25.699 --> 25.691 --> 25.671 --> 25.637 --> 25.585 --> 25.476 --> 25.350 --> 25.234 --> 25.144 --> 25.060 --> 24.980 --> 24.918 --> 24.878 --> 24.843 --> 24.821 --> 24.807 --> 24.794 --> 24.785 --> 24.778 --> 24.769 --> 24.757 --> 24.746 --> 24.733 --> 24.723 --> 24.711 --> 24.693 --> 24.671 --> 24.648 --> 24.624 --> 24.606 --> 24.594 --> 24.589 --> 24.587 --> 24.585 --> 24.583 --> 24.582 --> 24.579 --> 24.577 --> 24.574 --> 24.572 --> 24.570 --> 24.569 --> 24.568 --> 24.567 --> 24.566 --> 24.566 --> 24.565 --> 24.565 --> 24.565 --> 24.565 --> 24.565 --> 24.565 --> \n",
      "Init 13\n",
      "35.474 --> 25.483 --> 25.102 --> 25.006 --> 24.968 --> 24.947 --> 24.929 --> 24.914 --> 24.899 --> 24.875 --> 24.820 --> 24.691 --> 24.515 --> 24.419 --> 24.334 --> 24.270 --> 24.231 --> 24.205 --> 24.186 --> 24.175 --> 24.165 --> 24.151 --> 24.134 --> 24.121 --> 24.114 --> 24.111 --> 24.109 --> 24.109 --> 24.109 --> 24.109 --> 24.109 --> 24.108 --> 24.108 --> 24.108 --> 24.108 --> \n",
      "Init 15\n",
      "36.109 --> 25.615 --> 24.782 --> 24.420 --> 24.254 --> 24.198 --> 24.163 --> 24.141 --> 24.129 --> 24.124 --> 24.120 --> 24.119 --> 24.118 --> 24.117 --> 24.117 --> 24.116 --> 24.116 --> 24.116 --> 24.115 --> 24.115 --> 24.114 --> 24.113 --> 24.112 --> 24.111 --> 24.110 --> 24.109 --> 24.108 --> 24.108 --> 24.108 --> \n",
      "Init 17\n",
      "34.295 --> 25.686 --> 25.302 --> 25.149 --> 25.074 --> 25.036 --> 25.014 --> 24.998 --> 24.988 --> 24.972 --> 24.949 --> 24.933 --> 24.921 --> 24.910 --> 24.899 --> 24.886 --> 24.870 --> 24.851 --> 24.821 --> 24.711 --> 24.542 --> 24.436 --> 24.347 --> 24.276 --> 24.236 --> 24.211 --> 24.187 --> 24.176 --> 24.166 --> 24.153 --> 24.137 --> 24.123 --> 24.114 --> 24.111 --> 24.110 --> 24.109 --> 24.109 --> 24.109 --> 24.109 --> 24.108 --> 24.108 --> 24.108 --> 24.108 --> 24.108 --> \n",
      "Init 19\n",
      "34.954 --> 25.861 --> 25.397 --> 25.133 --> 24.986 --> 24.876 --> 24.825 --> 24.800 --> 24.786 --> 24.773 --> 24.759 --> 24.747 --> 24.732 --> 24.722 --> 24.710 --> 24.692 --> 24.670 --> 24.647 --> 24.624 --> 24.605 --> 24.594 --> 24.589 --> 24.587 --> 24.585 --> 24.583 --> 24.582 --> 24.580 --> 24.578 --> 24.574 --> 24.572 --> 24.570 --> 24.569 --> 24.568 --> 24.567 --> 24.566 --> 24.566 --> 24.565 --> 24.565 --> 24.565 --> 24.565 --> 24.565 --> 24.565 --> \n",
      "Init 21\n",
      "35.866 --> 26.286 --> 25.437 --> 25.111 --> 24.959 --> 24.889 --> 24.859 --> 24.840 --> 24.830 --> 24.825 --> 24.823 --> 24.820 --> 24.819 --> 24.818 --> 24.816 --> 24.814 --> 24.813 --> 24.812 --> 24.812 --> 24.812 --> 24.812 --> \n",
      "Init 23\n",
      "35.691 --> 25.404 --> 25.179 --> 25.068 --> 24.890 --> 24.641 --> 24.345 --> 24.244 --> 24.204 --> 24.186 --> 24.175 --> 24.168 --> 24.157 --> 24.142 --> 24.127 --> 24.116 --> 24.111 --> 24.110 --> 24.109 --> 24.109 --> 24.109 --> 24.109 --> 24.109 --> 24.109 --> 24.109 --> 24.109 --> 24.109 --> 24.109 --> 24.109 --> 24.109 --> \n",
      "Init 25\n",
      "31.996 --> 25.031 --> 24.737 --> 24.662 --> 24.635 --> 24.625 --> 24.623 --> 24.621 --> 24.618 --> 24.616 --> 24.616 --> 24.614 --> 24.612 --> 24.612 --> 24.611 --> 24.611 --> 24.611 --> 24.610 --> 24.610 --> 24.610 --> 24.610 --> 24.610 --> 24.610 --> \n",
      "Init 27\n",
      "36.067 --> 26.033 --> 24.982 --> 24.510 --> 24.320 --> 24.240 --> 24.211 --> 24.194 --> 24.186 --> 24.183 --> 24.180 --> 24.178 --> 24.175 --> 24.171 --> 24.161 --> 24.137 --> 24.093 --> 24.059 --> 24.047 --> 24.040 --> 24.035 --> 24.030 --> 24.025 --> 24.023 --> 24.021 --> 24.020 --> 24.018 --> 24.016 --> 24.015 --> 24.015 --> 24.014 --> 24.013 --> 24.013 --> 24.013 --> 24.013 --> 24.013 --> 24.013 --> 24.013 --> 24.012 --> 24.012 --> 24.012 --> 24.012 --> 24.012 --> 24.012 --> 24.012 --> 24.011 --> 24.011 --> 24.011 --> 24.011 --> 24.011 --> 24.011 --> 24.011 --> 24.011 --> 24.011 --> \n",
      "Init 29\n",
      "35.937 --> 25.837 --> 25.584 --> 25.520 --> 25.430 --> 25.287 --> 25.200 --> 25.122 --> 25.034 --> 24.992 --> 24.975 --> 24.965 --> 24.953 --> 24.943 --> 24.928 --> 24.914 --> 24.903 --> 24.890 --> 24.872 --> 24.854 --> 24.824 --> 24.719 --> 24.547 --> 24.439 --> 24.348 --> 24.276 --> 24.236 --> 24.211 --> 24.187 --> 24.176 --> 24.166 --> 24.153 --> 24.137 --> 24.123 --> 24.114 --> 24.111 --> 24.110 --> 24.109 --> 24.109 --> 24.109 --> 24.109 --> 24.108 --> 24.108 --> 24.108 --> 24.108 --> 24.108 --> \n",
      "Init 31\n",
      "33.314 --> 25.105 --> 24.911 --> 24.830 --> 24.770 --> 24.724 --> 24.700 --> 24.683 --> 24.678 --> 24.671 --> 24.663 --> 24.656 --> 24.651 --> 24.645 --> 24.640 --> 24.634 --> 24.628 --> 24.619 --> 24.609 --> 24.597 --> 24.589 --> 24.577 --> 24.561 --> 24.545 --> 24.533 --> 24.525 --> 24.517 --> 24.507 --> 24.496 --> 24.487 --> 24.484 --> 24.481 --> 24.479 --> 24.474 --> 24.471 --> 24.468 --> 24.464 --> 24.461 --> 24.460 --> 24.459 --> 24.459 --> 24.458 --> 24.457 --> 24.457 --> 24.457 --> 24.456 --> 24.456 --> 24.455 --> 24.455 --> 24.455 --> 24.454 --> 24.454 --> 24.454 --> 24.454 --> 24.454 --> 24.454 --> \n",
      "Init 33\n",
      "35.132 --> 25.930 --> 25.168 --> 24.789 --> 24.459 --> 24.245 --> 24.158 --> 24.132 --> 24.123 --> 24.119 --> 24.117 --> 24.115 --> 24.115 --> 24.115 --> 24.115 --> \n",
      "Init 35\n",
      "39.780 --> 24.673 --> 24.143 --> 24.042 --> 24.028 --> 24.023 --> 24.019 --> 24.016 --> 24.014 --> 24.012 --> 24.012 --> 24.012 --> 24.012 --> 24.012 --> 24.012 --> 24.012 --> 24.012 --> 24.012 --> 24.012 --> \n",
      "Init 37\n",
      "35.424 --> 25.378 --> 24.952 --> 24.694 --> 24.518 --> 24.404 --> 24.320 --> 24.252 --> 24.186 --> 24.145 --> 24.130 --> 24.122 --> 24.119 --> 24.118 --> 24.117 --> 24.117 --> 24.117 --> 24.117 --> 24.116 --> 24.116 --> 24.116 --> 24.116 --> 24.115 --> 24.115 --> 24.115 --> 24.114 --> 24.113 --> 24.112 --> 24.111 --> 24.110 --> 24.109 --> 24.108 --> 24.108 --> 24.108 --> \n",
      "Init 39\n",
      "34.143 --> 26.742 --> 24.959 --> 24.689 --> 24.644 --> 24.632 --> 24.625 --> 24.620 --> 24.618 --> 24.617 --> 24.616 --> 24.616 --> 24.615 --> 24.614 --> 24.614 --> 24.614 --> \n",
      "Init 41\n",
      "36.437 --> 25.933 --> 25.279 --> 25.046 --> 24.983 --> 24.936 --> 24.893 --> 24.846 --> 24.810 --> 24.780 --> 24.743 --> 24.707 --> 24.680 --> 24.651 --> 24.626 --> 24.609 --> 24.598 --> 24.591 --> 24.587 --> 24.584 --> 24.582 --> 24.580 --> 24.578 --> 24.574 --> 24.572 --> 24.570 --> 24.569 --> 24.568 --> 24.567 --> 24.566 --> 24.566 --> 24.565 --> 24.565 --> 24.565 --> 24.565 --> 24.565 --> 24.565 --> \n",
      "Init 43\n",
      "34.051 --> 26.401 --> 26.167 --> 26.048 --> 25.956 --> 25.917 --> 25.906 --> 25.901 --> 25.898 --> 25.897 --> 25.894 --> 25.893 --> 25.893 --> 25.893 --> \n",
      "Init 45\n",
      "30.350 --> 24.587 --> 24.329 --> 24.234 --> 24.181 --> 24.153 --> 24.137 --> 24.128 --> 24.122 --> 24.117 --> 24.116 --> 24.115 --> 24.115 --> 24.115 --> 24.114 --> 24.114 --> 24.114 --> 24.113 --> 24.112 --> 24.111 --> 24.109 --> 24.108 --> 24.107 --> 24.107 --> 24.107 --> 24.107 --> 24.107 --> \n",
      "Init 47\n",
      "36.365 --> 25.804 --> 24.971 --> 24.693 --> 24.548 --> 24.483 --> 24.440 --> 24.403 --> 24.369 --> 24.331 --> 24.288 --> 24.241 --> 24.184 --> 24.135 --> 24.099 --> 24.082 --> 24.071 --> 24.059 --> 24.047 --> 24.040 --> 24.034 --> 24.027 --> 24.021 --> 24.017 --> 24.015 --> 24.014 --> 24.013 --> 24.012 --> 24.012 --> 24.012 --> 24.012 --> 24.012 --> 24.012 --> 24.011 --> 24.011 --> 24.011 --> 24.011 --> 24.011 --> 24.011 --> 24.011 --> \n",
      "Init 49\n",
      "35.882 --> 25.390 --> 24.900 --> 24.715 --> 24.656 --> 24.635 --> 24.624 --> 24.617 --> 24.612 --> 24.611 --> 24.610 --> 24.610 --> 24.610 --> 24.610 --> 24.610 --> 24.610 --> \n",
      "Init 51\n",
      "43.128 --> 26.938 --> 26.378 --> 26.109 --> 25.736 --> 25.289 --> 25.034 --> 24.918 --> 24.868 --> 24.847 --> 24.839 --> 24.837 --> 24.837 --> 24.837 --> 24.837 --> \n",
      "Init 53\n",
      "34.117 --> 25.676 --> 25.328 --> 25.086 --> 24.843 --> 24.646 --> 24.495 --> 24.387 --> 24.309 --> 24.244 --> 24.184 --> 24.145 --> 24.129 --> 24.122 --> 24.120 --> 24.118 --> 24.117 --> 24.117 --> 24.117 --> 24.116 --> 24.116 --> 24.115 --> 24.115 --> 24.114 --> 24.113 --> 24.112 --> 24.111 --> 24.110 --> 24.109 --> 24.108 --> 24.108 --> 24.108 --> \n",
      "Init 55\n",
      "35.777 --> 25.514 --> 25.233 --> 25.151 --> 25.064 --> 25.006 --> 24.940 --> 24.898 --> 24.869 --> 24.830 --> 24.727 --> 24.552 --> 24.449 --> 24.367 --> 24.285 --> 24.241 --> 24.214 --> 24.192 --> 24.180 --> 24.170 --> 24.159 --> 24.143 --> 24.127 --> 24.116 --> 24.112 --> 24.110 --> 24.109 --> 24.109 --> 24.109 --> 24.109 --> 24.109 --> 24.108 --> 24.108 --> 24.108 --> 24.108 --> \n",
      "Init 57\n",
      "37.126 --> 25.976 --> 25.532 --> 25.103 --> 24.633 --> 24.335 --> 24.198 --> 24.122 --> 24.091 --> 24.073 --> 24.061 --> 24.048 --> 24.038 --> 24.031 --> 24.023 --> 24.018 --> 24.014 --> 24.012 --> 24.011 --> 24.011 --> 24.011 --> 24.011 --> 24.011 --> 24.011 --> 24.011 --> \n",
      "Init 59\n",
      "35.411 --> 24.754 --> 24.357 --> 24.258 --> 24.193 --> 24.141 --> 24.100 --> 24.082 --> 24.069 --> 24.057 --> 24.045 --> 24.036 --> 24.030 --> 24.024 --> 24.018 --> 24.014 --> 24.012 --> 24.011 --> 24.011 --> 24.011 --> 24.011 --> 24.011 --> 24.011 --> 24.011 --> \n",
      "Init 61\n",
      "40.185 --> 26.640 --> 26.455 --> 26.266 --> 25.646 --> 25.111 --> 24.945 --> 24.883 --> 24.854 --> 24.843 --> 24.839 --> 24.838 --> 24.837 --> 24.837 --> 24.837 --> 24.837 --> \n",
      "Init 63\n",
      "32.643 --> 24.651 --> 24.181 --> 24.098 --> 24.072 --> 24.057 --> 24.046 --> 24.035 --> 24.025 --> 24.019 --> 24.015 --> 24.013 --> 24.012 --> 24.011 --> 24.011 --> 24.011 --> 24.011 --> 24.011 --> \n",
      "Init 65\n",
      "36.987 --> 26.258 --> 25.229 --> 24.861 --> 24.660 --> 24.500 --> 24.372 --> 24.294 --> 24.225 --> 24.191 --> 24.169 --> 24.149 --> 24.132 --> 24.123 --> 24.119 --> 24.118 --> 24.117 --> 24.116 --> 24.115 --> 24.115 --> 24.113 --> 24.112 --> 24.110 --> 24.109 --> 24.108 --> 24.108 --> 24.108 --> \n",
      "Init 67\n",
      "36.448 --> 27.056 --> 26.656 --> 26.235 --> 25.590 --> 25.358 --> 25.274 --> 25.192 --> 25.055 --> 24.915 --> 24.840 --> 24.812 --> 24.796 --> 24.778 --> 24.763 --> 24.748 --> 24.734 --> 24.721 --> 24.704 --> 24.683 --> 24.661 --> 24.636 --> 24.614 --> 24.599 --> 24.592 --> 24.588 --> 24.586 --> 24.584 --> 24.582 --> 24.580 --> 24.578 --> 24.575 --> 24.572 --> 24.571 --> 24.569 --> 24.568 --> 24.567 --> 24.566 --> 24.566 --> 24.565 --> 24.565 --> 24.565 --> 24.565 --> 24.565 --> 24.565 --> \n",
      "Init 69\n",
      "33.242 --> 26.436 --> 25.798 --> 25.057 --> 24.869 --> 24.821 --> 24.787 --> 24.767 --> 24.749 --> 24.734 --> 24.710 --> 24.687 --> 24.661 --> 24.635 --> 24.613 --> 24.599 --> 24.592 --> 24.588 --> 24.585 --> 24.583 --> 24.581 --> 24.579 --> 24.577 --> 24.574 --> 24.572 --> 24.570 --> 24.569 --> 24.568 --> 24.567 --> 24.566 --> 24.566 --> 24.565 --> 24.565 --> 24.565 --> 24.565 --> 24.565 --> 24.565 --> \n",
      "Init 71\n",
      "34.636 --> 25.300 --> 24.923 --> 24.655 --> 24.472 --> 24.390 --> 24.316 --> 24.257 --> 24.221 --> 24.197 --> 24.182 --> 24.174 --> 24.165 --> 24.150 --> 24.131 --> 24.119 --> 24.113 --> 24.110 --> 24.109 --> 24.109 --> 24.109 --> 24.109 --> 24.108 --> 24.108 --> 24.108 --> 24.108 --> 24.108 --> \n",
      "Init 73\n",
      "41.475 --> 27.367 --> 26.386 --> 25.404 --> 24.908 --> 24.717 --> 24.594 --> 24.506 --> 24.452 --> 24.393 --> 24.315 --> 24.234 --> 24.176 --> 24.138 --> 24.128 --> 24.124 --> 24.121 --> 24.118 --> 24.117 --> 24.117 --> 24.116 --> 24.116 --> 24.116 --> 24.116 --> 24.116 --> 24.116 --> 24.116 --> \n",
      "Init 75\n",
      "31.835 --> 24.996 --> 24.577 --> 24.454 --> 24.331 --> 24.251 --> 24.206 --> 24.179 --> 24.159 --> 24.147 --> 24.142 --> 24.137 --> 24.131 --> 24.121 --> 24.115 --> 24.113 --> 24.111 --> 24.110 --> 24.109 --> 24.108 --> 24.107 --> 24.107 --> 24.107 --> 24.107 --> \n",
      "Init 77\n",
      "37.564 --> 25.992 --> 25.333 --> 24.974 --> 24.868 --> 24.830 --> 24.810 --> 24.798 --> 24.787 --> 24.781 --> 24.774 --> 24.765 --> 24.754 --> 24.742 --> 24.729 --> 24.717 --> 24.701 --> 24.680 --> 24.655 --> 24.631 --> 24.611 --> 24.598 --> 24.591 --> 24.587 --> 24.585 --> 24.583 --> 24.582 --> 24.580 --> 24.578 --> 24.575 --> 24.572 --> 24.571 --> 24.569 --> 24.568 --> 24.567 --> 24.566 --> 24.566 --> 24.565 --> 24.565 --> 24.565 --> 24.565 --> 24.565 --> 24.565 --> \n",
      "Init 79\n",
      "37.942 --> 25.110 --> 24.868 --> 24.830 --> 24.821 --> 24.815 --> 24.813 --> 24.812 --> 24.812 --> 24.812 --> 24.812 --> 24.812 --> \n",
      "Init 81\n",
      "34.761 --> 27.148 --> 26.308 --> 25.888 --> 25.703 --> 25.557 --> 25.431 --> 25.306 --> 25.196 --> 25.058 --> 24.864 --> 24.720 --> 24.643 --> 24.598 --> 24.582 --> 24.573 --> 24.568 --> 24.566 --> 24.565 --> 24.565 --> 24.565 --> 24.565 --> 24.565 --> \n",
      "Init 83\n",
      "34.503 --> 24.416 --> 24.231 --> 24.184 --> 24.164 --> 24.153 --> 24.145 --> 24.139 --> 24.132 --> 24.125 --> 24.120 --> 24.117 --> 24.114 --> 24.111 --> 24.109 --> 24.108 --> 24.107 --> 24.107 --> 24.107 --> 24.107 --> \n",
      "Init 85\n",
      "36.234 --> 26.059 --> 25.199 --> 24.904 --> 24.753 --> 24.686 --> 24.660 --> 24.647 --> 24.641 --> 24.637 --> 24.632 --> 24.630 --> 24.629 --> 24.628 --> 24.627 --> 24.626 --> 24.624 --> 24.624 --> 24.624 --> 24.624 --> \n",
      "Init 87\n",
      "33.049 --> 24.671 --> 24.418 --> 24.385 --> 24.379 --> 24.377 --> 24.376 --> 24.376 --> 24.375 --> 24.375 --> 24.375 --> 24.375 --> 24.375 --> 24.375 --> 24.375 --> \n",
      "Init 89\n",
      "37.180 --> 25.466 --> 24.907 --> 24.832 --> 24.797 --> 24.780 --> 24.769 --> 24.756 --> 24.743 --> 24.731 --> 24.721 --> 24.708 --> 24.690 --> 24.670 --> 24.647 --> 24.623 --> 24.605 --> 24.594 --> 24.589 --> 24.587 --> 24.585 --> 24.583 --> 24.582 --> 24.580 --> 24.578 --> 24.574 --> 24.572 --> 24.570 --> 24.569 --> 24.568 --> 24.567 --> 24.566 --> 24.566 --> 24.565 --> 24.565 --> 24.565 --> 24.565 --> 24.565 --> 24.565 --> \n",
      "Init 91\n",
      "37.233 --> 25.446 --> 25.136 --> 24.957 --> 24.840 --> 24.771 --> 24.723 --> 24.693 --> 24.668 --> 24.643 --> 24.624 --> 24.609 --> 24.600 --> 24.595 --> 24.590 --> 24.583 --> 24.581 --> 24.580 --> 24.579 --> 24.578 --> 24.578 --> 24.578 --> 24.578 --> 24.577 --> 24.577 --> 24.576 --> 24.576 --> 24.576 --> 24.576 --> 24.576 --> \n",
      "Init 93\n",
      "34.869 --> 26.043 --> 25.584 --> 25.256 --> 25.096 --> 24.999 --> 24.907 --> 24.819 --> 24.745 --> 24.685 --> 24.650 --> 24.621 --> 24.601 --> 24.590 --> 24.579 --> 24.571 --> 24.567 --> 24.564 --> 24.563 --> 24.563 --> 24.562 --> 24.562 --> 24.562 --> 24.562 --> 24.562 --> 24.562 --> \n",
      "Init 95\n",
      "35.529 --> 25.838 --> 25.306 --> 25.095 --> 24.974 --> 24.897 --> 24.846 --> 24.824 --> 24.808 --> 24.796 --> 24.786 --> 24.779 --> 24.771 --> 24.760 --> 24.749 --> 24.735 --> 24.724 --> 24.711 --> 24.693 --> 24.671 --> 24.648 --> 24.624 --> 24.606 --> 24.594 --> 24.589 --> 24.587 --> 24.585 --> 24.583 --> 24.582 --> 24.580 --> 24.578 --> 24.574 --> 24.572 --> 24.570 --> 24.569 --> 24.568 --> 24.567 --> 24.566 --> 24.566 --> 24.565 --> 24.565 --> 24.565 --> 24.565 --> 24.565 --> 24.565 --> \n",
      "Init 97\n",
      "31.995 --> 26.409 --> 26.223 --> 26.084 --> 25.908 --> 25.709 --> 25.519 --> 25.283 --> 25.122 --> 25.002 --> 24.931 --> 24.881 --> 24.845 --> 24.824 --> 24.810 --> 24.796 --> 24.785 --> 24.778 --> 24.768 --> 24.756 --> 24.745 --> 24.732 --> 24.722 --> 24.710 --> 24.692 --> 24.671 --> 24.648 --> 24.624 --> 24.606 --> 24.594 --> 24.589 --> 24.587 --> 24.585 --> 24.583 --> 24.582 --> 24.579 --> 24.577 --> 24.574 --> 24.572 --> 24.570 --> 24.569 --> 24.568 --> 24.567 --> 24.566 --> 24.566 --> 24.565 --> 24.565 --> 24.565 --> 24.565 --> 24.565 --> 24.565 --> \n",
      "Init 99\n",
      "31.104 --> 25.217 --> 24.696 --> 24.446 --> 24.321 --> 24.254 --> 24.212 --> 24.193 --> 24.183 --> 24.175 --> 24.164 --> 24.150 --> 24.131 --> 24.118 --> 24.113 --> 24.110 --> 24.109 --> 24.109 --> 24.109 --> 24.109 --> 24.108 --> 24.108 --> 24.108 --> 24.108 --> 24.108 --> \n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "TimeSeriesKMeans(dtw_inertia=False, max_iter=100, max_iter_barycenter=100,\n",
       "         metric='euclidean', metric_params=None, n_clusters=5, n_init=100,\n",
       "         random_state=2019, tol=1e-06, verbose=1)"
      ]
     },
     "execution_count": 54,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Train using Time Series k-Means\n",
    "km = TimeSeriesKMeans(n_clusters=5, max_iter=100, n_init=100, \\\n",
    "                      metric=\"euclidean\", verbose=1, random_state=2019)\n",
    "km.fit(X_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 55,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Adjusted Rand Index of Time Series k-Means on Training Set: 0.5063464656715959\n"
     ]
    }
   ],
   "source": [
    "# Predict on training set and evaluate using adjusted Rand index\n",
    "preds = km.predict(X_train)\n",
    "ars = adjusted_rand_score(data_train[:,0],preds)\n",
    "print(\"Adjusted Rand Index of Time Series k-Means on Training Set:\", ars)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 56,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Adjusted Rand Index of Time Series k-Means on Test Set: 0.4864981997585834\n"
     ]
    }
   ],
   "source": [
    "# Predict on test set and evaluate using adjusted Rand index\n",
    "preds_test = km.predict(X_test)\n",
    "ars = adjusted_rand_score(data_test[:,0],preds_test)\n",
    "print(\"Adjusted Rand Index of Time Series k-Means on Test Set:\", ars)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 60,
   "metadata": {
    "collapsed": true,
    "jupyter": {
     "outputs_hidden": true
    }
   },
   "outputs": [],
   "source": [
    "# HDBSCAN"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 61,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Adjusted Rand Index on Training Set: 0.7689563655060421\n"
     ]
    }
   ],
   "source": [
    "# Train model and evaluate on training set\n",
    "min_cluster_size = 5\n",
    "min_samples = None\n",
    "alpha = 1.0\n",
    "cluster_selection_method = 'eom'\n",
    "prediction_data = True\n",
    "\n",
    "hdb = hdbscan.HDBSCAN(min_cluster_size=min_cluster_size, \\\n",
    "                      min_samples=min_samples, alpha=alpha, \\\n",
    "                      cluster_selection_method=cluster_selection_method, \\\n",
    "                      prediction_data=prediction_data)\n",
    "\n",
    "preds = hdb.fit_predict(X_train.reshape(4000,140))\n",
    "ars = adjusted_rand_score(data_train[:,0],preds)\n",
    "print(\"Adjusted Rand Index on Training Set:\", ars)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 62,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Adjusted Rand Index on Test Set: 0.7200816245545564\n"
     ]
    }
   ],
   "source": [
    "# Predict on test set and evaluate\n",
    "preds_test = hdbscan.prediction.approximate_predict( \\\n",
    "                hdb, X_test.reshape(1000,140))\n",
    "ars = adjusted_rand_score(data_test[:,0],preds_test[0])\n",
    "print(\"Adjusted Rand Index on Test Set:\", ars)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true,
    "jupyter": {
     "outputs_hidden": true
    }
   },
   "outputs": [],
   "source": [
    "# Experiment to compare time series clustering algorithms"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 63,
   "metadata": {
    "collapsed": true,
    "jupyter": {
     "outputs_hidden": true
    }
   },
   "outputs": [],
   "source": [
    "# Load the datasets\n",
    "current_path = os.getcwd()\n",
    "file = os.path.sep.join(['', 'datasets', 'ucr_time_series_data', ''])\n",
    "\n",
    "mypath = current_path + file\n",
    "d = []\n",
    "f = []\n",
    "for (dirpath, dirnames, filenames) in walk(mypath):\n",
    "    for i in dirnames:\n",
    "        newpath = os.path.sep.join([mypath, i, \"\"])\n",
    "        onlyfiles = [f for f in listdir(newpath) if isfile(join(newpath, f))]\n",
    "        f.extend(onlyfiles)\n",
    "    d.extend(dirnames)\n",
    "    break"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true,
    "jupyter": {
     "outputs_hidden": true
    }
   },
   "outputs": [],
   "source": [
    "# k-Shape Experiment - FULL RUN"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 68,
   "metadata": {
    "collapsed": true,
    "jupyter": {
     "outputs_hidden": true
    }
   },
   "outputs": [],
   "source": [
    "# Create dataframe\n",
    "kShapeDF = pd.DataFrame(data=[],index=[v for v in d], \\\n",
    "                        columns=[\"Train ARS\",\"Test ARS\"])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 69,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Dataset  1\n",
      "50words\n",
      "Adjusted Rand Index on Training Set: 0.33141709829469923\n",
      "Adjusted Rand Index on Test Set: 0.28178766416273154\n",
      "Dataset  2\n",
      "Adiac\n",
      "Adjusted Rand Index on Training Set: 0.0056982414737593275\n",
      "Adjusted Rand Index on Test Set: 0.0032752154154245226\n",
      "Dataset  3\n",
      "ArrowHead\n",
      "Adjusted Rand Index on Training Set: 0.1171844628000625\n",
      "Adjusted Rand Index on Test Set: 0.030556785699907458\n",
      "Dataset  4\n",
      "Beef\n",
      "Adjusted Rand Index on Training Set: 0.07754088914498472\n",
      "Adjusted Rand Index on Test Set: -0.08641975308641976\n",
      "Dataset  5\n",
      "BeetleFly\n",
      "Adjusted Rand Index on Training Set: 0.03236819430670938\n",
      "Adjusted Rand Index on Test Set: -0.1487179487179487\n",
      "Dataset  6\n",
      "BirdChicken\n",
      "Adjusted Rand Index on Training Set: -0.007756188019027647\n",
      "Adjusted Rand Index on Test Set: 0.0\n",
      "Dataset  7\n",
      "Car\n",
      "Adjusted Rand Index on Training Set: 0.08302004569779552\n",
      "Adjusted Rand Index on Test Set: 0.11675289348765265\n",
      "Dataset  8\n",
      "CBF\n",
      "Adjusted Rand Index on Training Set: 0.4300191679141888\n",
      "Adjusted Rand Index on Test Set: 0.4464515716637103\n",
      "Dataset  9\n",
      "ChlorineConcentration\n",
      "Adjusted Rand Index on Training Set: 0.004417834137649424\n",
      "Adjusted Rand Index on Test Set: 0.0010502228635036084\n",
      "Dataset  10\n",
      "CinC_ECG_torso\n",
      "Adjusted Rand Index on Training Set: 0.07495613778505084\n",
      "Adjusted Rand Index on Test Set: 0.08140683701809101\n",
      "Dataset  11\n",
      "Coffee\n",
      "Adjusted Rand Index on Training Set: -0.010697146691743027\n",
      "Adjusted Rand Index on Test Set: 0.0638297872340425\n",
      "Dataset  12\n",
      "Computers\n",
      "Adjusted Rand Index on Training Set: 0.023161070057011277\n",
      "Adjusted Rand Index on Test Set: 0.13575757575757577\n",
      "Dataset  13\n",
      "Cricket_X\n",
      "Adjusted Rand Index on Training Set: 0.16976146445673967\n",
      "Adjusted Rand Index on Test Set: 0.182969943211159\n",
      "Dataset  14\n",
      "Cricket_Y\n",
      "Adjusted Rand Index on Training Set: 0.22286240401119306\n",
      "Adjusted Rand Index on Test Set: 0.23224160887267567\n",
      "Dataset  15\n",
      "Cricket_Z\n",
      "Adjusted Rand Index on Training Set: 0.16526403018829078\n",
      "Adjusted Rand Index on Test Set: 0.15001695403251467\n",
      "Dataset  16\n",
      "DiatomSizeReduction\n",
      "Adjusted Rand Index on Training Set: 0.015823817916421807\n",
      "Adjusted Rand Index on Test Set: 0.0\n",
      "Dataset  17\n",
      "DistalPhalanxOutlineAgeGroup\n",
      "Adjusted Rand Index on Training Set: 0.42341995617548944\n",
      "Adjusted Rand Index on Test Set: 0.32791136957547806\n",
      "Dataset  18\n",
      "DistalPhalanxOutlineCorrect\n",
      "Adjusted Rand Index on Training Set: 0.01160268699565799\n",
      "Adjusted Rand Index on Test Set: 0.0\n",
      "Dataset  19\n",
      "DistalPhalanxTW\n",
      "Adjusted Rand Index on Training Set: 0.002928701689761342\n",
      "Adjusted Rand Index on Test Set: 3.607080699415818e-05\n",
      "Dataset  20\n",
      "Earthquakes\n",
      "Adjusted Rand Index on Training Set: 0.0037961924568304505\n",
      "Adjusted Rand Index on Test Set: -0.002443897563716349\n",
      "Dataset  21\n",
      "ECG200\n",
      "Adjusted Rand Index on Training Set: 0.29496493777244304\n",
      "Adjusted Rand Index on Test Set: 0.10138248847926272\n",
      "Dataset  22\n",
      "ECG5000\n",
      "Adjusted Rand Index on Training Set: 0.5463682297549879\n",
      "Adjusted Rand Index on Test Set: 0.5506884913514337\n",
      "Dataset  23\n",
      "ECGFiveDays\n",
      "Adjusted Rand Index on Training Set: 0.7367521607772766\n",
      "Adjusted Rand Index on Test Set: 0.6879858106122982\n",
      "Dataset  24\n",
      "ElectricDevices\n",
      "Adjusted Rand Index on Training Set: 0.10408716136327975\n",
      "Adjusted Rand Index on Test Set: 0.11337296287421782\n",
      "Dataset  25\n",
      "FaceAll\n",
      "Adjusted Rand Index on Training Set: 0.32492880780839567\n",
      "Adjusted Rand Index on Test Set: 0.35397645785525017\n",
      "Dataset  26\n",
      "FaceFour\n",
      "Adjusted Rand Index on Training Set: 0.4557529657687818\n",
      "Adjusted Rand Index on Test Set: 0.33962549650085117\n",
      "Dataset  27\n",
      "FacesUCR\n",
      "Adjusted Rand Index on Training Set: 0.39001583238124987\n",
      "Adjusted Rand Index on Test Set: 0.34940154851786337\n",
      "Dataset  28\n",
      "FISH\n",
      "Adjusted Rand Index on Training Set: 0.0033148293387528344\n",
      "Adjusted Rand Index on Test Set: -0.004518505675836564\n",
      "Dataset  29\n",
      "FordA\n",
      "Adjusted Rand Index on Training Set: 0.13549231655358876\n",
      "Adjusted Rand Index on Test Set: 0.11615890520504485\n",
      "Dataset  30\n",
      "FordB\n",
      "Adjusted Rand Index on Training Set: 0.02688237531510243\n",
      "Adjusted Rand Index on Test Set: 0.03370268756111131\n",
      "Dataset  31\n",
      "Gun_Point\n",
      "Adjusted Rand Index on Training Set: -0.004912969824990243\n",
      "Adjusted Rand Index on Test Set: -0.003157686994276681\n",
      "Dataset  32\n",
      "Ham\n",
      "Adjusted Rand Index on Training Set: 0.0837497989377934\n",
      "Adjusted Rand Index on Test Set: 0.10028860028860029\n",
      "Dataset  33\n",
      "HandOutlines\n",
      "Adjusted Rand Index on Training Set: 0.01932216693971623\n",
      "Adjusted Rand Index on Test Set: 0.009727348439364365\n",
      "Dataset  34\n",
      "Haptics\n",
      "Adjusted Rand Index on Training Set: 0.00014352853224834877\n",
      "Adjusted Rand Index on Test Set: 0.0\n",
      "Dataset  35\n",
      "Herring\n",
      "Adjusted Rand Index on Training Set: 0.010263861988621724\n",
      "Adjusted Rand Index on Test Set: -0.0907783282723348\n",
      "Dataset  36\n",
      "InlineSkate\n",
      "Adjusted Rand Index on Training Set: 0.026595569163195637\n",
      "Adjusted Rand Index on Test Set: 0.03416886789361377\n",
      "Dataset  37\n",
      "InsectWingbeatSound\n",
      "Adjusted Rand Index on Training Set: 0.2503909077282778\n",
      "Adjusted Rand Index on Test Set: 0.223498081090469\n",
      "Dataset  38\n",
      "ItalyPowerDemand\n",
      "Adjusted Rand Index on Training Set: 0.0006759158259042806\n",
      "Adjusted Rand Index on Test Set: -0.0015630592113879673\n",
      "Dataset  39\n",
      "LargeKitchenAppliances\n",
      "Adjusted Rand Index on Training Set: 0.16222135656274403\n",
      "Adjusted Rand Index on Test Set: 0.10254712730235327\n",
      "Dataset  40\n",
      "Lighting2\n",
      "Adjusted Rand Index on Training Set: 0.061208336997625463\n",
      "Adjusted Rand Index on Test Set: 0.0384615384615384\n",
      "Dataset  41\n",
      "Lighting7\n",
      "Adjusted Rand Index on Training Set: 0.31197880656040405\n",
      "Adjusted Rand Index on Test Set: 0.5092794759825328\n",
      "Dataset  42\n",
      "MALLAT\n",
      "Adjusted Rand Index on Training Set: 0.1321934048613034\n",
      "Adjusted Rand Index on Test Set: 0.12902334138718327\n",
      "Dataset  43\n",
      "Meat\n",
      "Adjusted Rand Index on Training Set: 0.0\n",
      "Adjusted Rand Index on Test Set: 0.0\n",
      "Dataset  44\n",
      "MedicalImages\n",
      "Adjusted Rand Index on Training Set: 0.047129541149609794\n",
      "Adjusted Rand Index on Test Set: 0.09816318178745811\n",
      "Dataset  45\n",
      "MiddlePhalanxOutlineAgeGroup\n",
      "Adjusted Rand Index on Training Set: 0.053061395625242105\n",
      "Adjusted Rand Index on Test Set: 0.08829908586059725\n",
      "Dataset  46\n",
      "MiddlePhalanxOutlineCorrect\n",
      "Adjusted Rand Index on Training Set: -0.0010721768145707043\n",
      "Adjusted Rand Index on Test Set: 0.0\n",
      "Dataset  47\n",
      "MiddlePhalanxTW\n",
      "Adjusted Rand Index on Training Set: 0.02789259905128924\n",
      "Adjusted Rand Index on Test Set: 0.011206350001384029\n",
      "Dataset  48\n",
      "MoteStrain\n",
      "Adjusted Rand Index on Training Set: 0.607571338768688\n",
      "Adjusted Rand Index on Test Set: 0.5952522880415954\n",
      "Dataset  49\n",
      "NonInvasiveFatalECG_Thorax1\n",
      "Adjusted Rand Index on Training Set: 0.02261835868266628\n",
      "Adjusted Rand Index on Test Set: 0.022489848009247908\n",
      "Dataset  50\n",
      "NonInvasiveFatalECG_Thorax2\n",
      "Adjusted Rand Index on Training Set: 0.013020100814681966\n",
      "Adjusted Rand Index on Test Set: 0.0140355935600567\n",
      "Dataset  51\n",
      "OliveOil\n",
      "Adjusted Rand Index on Training Set: 0.0\n",
      "Adjusted Rand Index on Test Set: 0.0\n",
      "Dataset  52\n",
      "OSULeaf\n",
      "Adjusted Rand Index on Training Set: 0.2696371023235093\n",
      "Adjusted Rand Index on Test Set: 0.2891733824792923\n",
      "Dataset  53\n",
      "PhalangesOutlinesCorrect\n",
      "Adjusted Rand Index on Training Set: 0.007880445352458241\n",
      "Adjusted Rand Index on Test Set: 0.004509857256984513\n",
      "Dataset  54\n",
      "Phoneme\n",
      "Adjusted Rand Index on Training Set: 0.03766196574544099\n",
      "Adjusted Rand Index on Test Set: 0.04559668943506461\n",
      "Dataset  55\n",
      "Plane\n",
      "Adjusted Rand Index on Training Set: 0.5022319842539696\n",
      "Adjusted Rand Index on Test Set: 0.437205603886871\n",
      "Dataset  56\n",
      "ProximalPhalanxOutlineAgeGroup\n",
      "Adjusted Rand Index on Training Set: 0.029132901055327596\n",
      "Adjusted Rand Index on Test Set: 0.15292065601630644\n",
      "Dataset  57\n",
      "ProximalPhalanxOutlineCorrect\n",
      "Adjusted Rand Index on Training Set: 0.003189915370497478\n",
      "Adjusted Rand Index on Test Set: 0.0\n",
      "Dataset  58\n",
      "ProximalPhalanxTW\n",
      "Adjusted Rand Index on Training Set: 0.11042555894629481\n",
      "Adjusted Rand Index on Test Set: 0.1417941905568511\n",
      "Dataset  59\n",
      "RefrigerationDevices\n",
      "Adjusted Rand Index on Training Set: 0.008796213532723463\n",
      "Adjusted Rand Index on Test Set: -0.006772588135974769\n",
      "Dataset  60\n",
      "ScreenType\n",
      "Adjusted Rand Index on Training Set: 0.014205397273558722\n",
      "Adjusted Rand Index on Test Set: -0.003332502170846456\n",
      "Dataset  61\n",
      "ShapeletSim\n",
      "Adjusted Rand Index on Training Set: 0.5226668101133044\n",
      "Adjusted Rand Index on Test Set: 0.13461538461538458\n",
      "Dataset  62\n",
      "ShapesAll\n",
      "Adjusted Rand Index on Training Set: 0.20179282447778674\n",
      "Adjusted Rand Index on Test Set: 0.17703951455473832\n",
      "Dataset  63\n",
      "SmallKitchenAppliances\n",
      "Adjusted Rand Index on Training Set: 0.023407702303042142\n",
      "Adjusted Rand Index on Test Set: 0.015610823173483574\n",
      "Dataset  64\n",
      "SonyAIBORobotSurface\n",
      "Adjusted Rand Index on Training Set: 0.047769307656884974\n",
      "Adjusted Rand Index on Test Set: 0.10178929994721742\n",
      "Dataset  65\n",
      "SonyAIBORobotSurfaceII\n",
      "Adjusted Rand Index on Training Set: 0.34388079746730893\n",
      "Adjusted Rand Index on Test Set: 0.3415584415584415\n",
      "Dataset  66\n",
      "StarLightCurves\n",
      "Adjusted Rand Index on Training Set: 0.49834088020580236\n",
      "Adjusted Rand Index on Test Set: 0.47949777351601575\n",
      "Dataset  67\n",
      "Strawberry\n",
      "Adjusted Rand Index on Training Set: -0.004303367134402091\n",
      "Adjusted Rand Index on Test Set: 0.0\n",
      "Dataset  68\n",
      "SwedishLeaf\n",
      "Adjusted Rand Index on Training Set: 0.11856009198162196\n",
      "Adjusted Rand Index on Test Set: 0.1209977738179147\n",
      "Dataset  69\n",
      "Symbols\n",
      "Adjusted Rand Index on Training Set: 0.7611708849510274\n",
      "Adjusted Rand Index on Test Set: 0.8139995755444789\n",
      "Dataset  70\n",
      "synthetic_control\n",
      "Adjusted Rand Index on Training Set: 0.5800018740031629\n",
      "Adjusted Rand Index on Test Set: 0.5351288659793815\n",
      "Dataset  71\n",
      "ToeSegmentation1\n",
      "Adjusted Rand Index on Training Set: -0.004374586372759009\n",
      "Adjusted Rand Index on Test Set: 0.0035159580546204423\n",
      "Dataset  72\n",
      "ToeSegmentation2\n",
      "Adjusted Rand Index on Training Set: 0.27233885527651797\n",
      "Adjusted Rand Index on Test Set: 0.091762492561095\n",
      "Dataset  73\n",
      "Trace\n",
      "Adjusted Rand Index on Training Set: 0.3614864785448748\n",
      "Adjusted Rand Index on Test Set: 0.3928223114984387\n",
      "Dataset  74\n",
      "TwoLeadECG\n",
      "Adjusted Rand Index on Training Set: 0.10410943422550015\n",
      "Adjusted Rand Index on Test Set: 0.06471155871935959\n",
      "Dataset  75\n",
      "Two_Patterns\n",
      "Adjusted Rand Index on Training Set: 0.24607567135896274\n",
      "Adjusted Rand Index on Test Set: 0.2691930999654907\n",
      "Dataset  76\n",
      "UWaveGestureLibraryAll\n",
      "Adjusted Rand Index on Training Set: 0.612735941277658\n",
      "Adjusted Rand Index on Test Set: 0.6152910803602317\n",
      "Dataset  77\n",
      "uWaveGestureLibrary_X\n",
      "Adjusted Rand Index on Training Set: 0.3464331818971397\n",
      "Adjusted Rand Index on Test Set: 0.34571916120112384\n",
      "Dataset  78\n",
      "uWaveGestureLibrary_Y\n",
      "Adjusted Rand Index on Training Set: 0.2722758749082347\n",
      "Adjusted Rand Index on Test Set: 0.25131476716649814\n",
      "Dataset  79\n",
      "uWaveGestureLibrary_Z\n",
      "Adjusted Rand Index on Training Set: 0.40111833372162314\n",
      "Adjusted Rand Index on Test Set: 0.3990591553178942\n",
      "Dataset  80\n",
      "wafer\n",
      "Adjusted Rand Index on Training Set: 0.06896163835273661\n",
      "Adjusted Rand Index on Test Set: 0.08666606446352638\n",
      "Dataset  81\n",
      "Wine\n",
      "Adjusted Rand Index on Training Set: -0.0065497213912263305\n",
      "Adjusted Rand Index on Test Set: -0.028905712319339343\n",
      "Dataset  82\n",
      "WordsSynonyms\n",
      "Adjusted Rand Index on Training Set: 0.21360548498006937\n",
      "Adjusted Rand Index on Test Set: 0.19426099492598015\n",
      "Dataset  83\n",
      "Worms\n",
      "Adjusted Rand Index on Training Set: 0.04981357783082403\n",
      "Adjusted Rand Index on Test Set: 0.053999156861993636\n",
      "Dataset  84\n",
      "WormsTwoClass\n",
      "Adjusted Rand Index on Training Set: 0.004658376142880903\n",
      "Adjusted Rand Index on Test Set: -0.013097576948264524\n",
      "Dataset  85\n",
      "yoga\n",
      "Adjusted Rand Index on Training Set: -0.0017217631275397603\n",
      "Adjusted Rand Index on Test Set: -0.003091207513006575\n",
      "Elapsed: 1.2039239147636625 hr \n"
     ]
    }
   ],
   "source": [
    "# Train and Evaluate k-Shape\n",
    "class ElapsedTimer(object):\n",
    "    def __init__(self):\n",
    "        self.start_time = time.time()\n",
    "    def elapsed(self,sec):\n",
    "        if sec < 60:\n",
    "            return str(sec) + \" sec\"\n",
    "        elif sec < (60 * 60):\n",
    "            return str(sec / 60) + \" min\"\n",
    "        else:\n",
    "            return str(sec / (60 * 60)) + \" hr\"\n",
    "    def elapsed_time(self):\n",
    "        print(\"Elapsed: %s \" % self.elapsed(time.time() - self.start_time))\n",
    "        return (time.time() - self.start_time)\n",
    "\n",
    "timer = ElapsedTimer()\n",
    "cnt = 0\n",
    "for i in d:\n",
    "    cnt += 1\n",
    "    print(\"Dataset \", cnt)\n",
    "    newpath = os.path.sep.join([mypath, i, \"\"])\n",
    "    onlyfiles = [f for f in listdir(newpath) if isfile(join(newpath, f))]\n",
    "    j = onlyfiles[0]\n",
    "    k = onlyfiles[1]\n",
    "    data_train = np.loadtxt(newpath+j, delimiter=\",\")\n",
    "    data_test = np.loadtxt(newpath+k, delimiter=\",\")\n",
    "\n",
    "    data_joined = np.concatenate((data_train,data_test),axis=0)\n",
    "    data_train, data_test = train_test_split(data_joined, \\\n",
    "                                        test_size=0.20, random_state=2019)\n",
    "\n",
    "    X_train = to_time_series_dataset(data_train[:, 1:])\n",
    "    y_train = data_train[:, 0].astype(np.int)\n",
    "    X_test = to_time_series_dataset(data_test[:, 1:])\n",
    "    y_test = data_test[:, 0].astype(np.int)\n",
    "    \n",
    "    X_train = TimeSeriesScalerMeanVariance(mu=0., std=1.) \\\n",
    "                                .fit_transform(X_train)\n",
    "    X_test = TimeSeriesScalerMeanVariance(mu=0., std=1.) \\\n",
    "                                .fit_transform(X_test)\n",
    "    \n",
    "    classes = len(np.unique(data_train[:,0]))\n",
    "    ks = KShape(n_clusters=classes, max_iter=10, n_init=3,verbose=0)\n",
    "    ks.fit(X_train)\n",
    "\n",
    "    print(i)\n",
    "    preds = ks.predict(X_train)\n",
    "    ars = adjusted_rand_score(data_train[:,0],preds)\n",
    "    print(\"Adjusted Rand Index on Training Set:\", ars)\n",
    "    kShapeDF.loc[i,\"Train ARS\"] = ars\n",
    "    \n",
    "    preds_test = ks.predict(X_test)\n",
    "    ars = adjusted_rand_score(data_test[:,0],preds_test)\n",
    "    print(\"Adjusted Rand Index on Test Set:\", ars)\n",
    "    kShapeDF.loc[i,\"Test ARS\"] = ars\n",
    "    \n",
    "kShapeTime = timer.elapsed_time()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 71,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Time to Run k-Shape Experiment in Minutes: 72.23543488581976\n"
     ]
    }
   ],
   "source": [
    "print(\"Time to Run k-Shape Experiment in Minutes:\", kShapeTime/60)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 73,
   "metadata": {
    "collapsed": true,
    "jupyter": {
     "outputs_hidden": true
    }
   },
   "outputs": [],
   "source": [
    "kShapeDF.to_pickle(os.path.sep.join([current_path, 'datasets', 'pickled_data', \\\n",
    "                   \"kShapeDF.pickle\"]))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 74,
   "metadata": {
    "collapsed": true,
    "jupyter": {
     "outputs_hidden": true
    }
   },
   "outputs": [],
   "source": [
    "# k-Means Experiment - FULL RUN"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 75,
   "metadata": {
    "collapsed": true,
    "jupyter": {
     "outputs_hidden": true
    }
   },
   "outputs": [],
   "source": [
    "# Create dataframe\n",
    "kMeansDF = pd.DataFrame(data=[],index=[v for v in d], \\\n",
    "                        columns=[\"Train ARS\",\"Test ARS\"])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 76,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Dataset  1\n",
      "50words\n",
      "Adjusted Rand Index on Training Set: 0.10619130288518851\n",
      "Adjusted Rand Index on Test Set: 0.08322258807014249\n",
      "Dataset  2\n",
      "Adiac\n",
      "Adjusted Rand Index on Training Set: 0.032298974657873764\n",
      "Adjusted Rand Index on Test Set: 0.03370617406524666\n",
      "Dataset  3\n",
      "ArrowHead\n",
      "Adjusted Rand Index on Training Set: 0.192471764199164\n",
      "Adjusted Rand Index on Test Set: 0.14002053273698203\n",
      "Dataset  4\n",
      "Beef\n",
      "Adjusted Rand Index on Training Set: 0.05051344138284522\n",
      "Adjusted Rand Index on Test Set: 0.09465020576131687\n",
      "Dataset  5\n",
      "BeetleFly\n",
      "Adjusted Rand Index on Training Set: 0.04482941020893944\n",
      "Adjusted Rand Index on Test Set: 0.17204301075268819\n",
      "Dataset  6\n",
      "BirdChicken\n",
      "Adjusted Rand Index on Training Set: 0.08246758609298872\n",
      "Adjusted Rand Index on Test Set: -0.026178010471204227\n",
      "Dataset  7\n",
      "Car\n",
      "Adjusted Rand Index on Training Set: 0.18277640411168794\n",
      "Adjusted Rand Index on Test Set: 0.048341493690090366\n",
      "Dataset  8\n",
      "CBF\n",
      "Adjusted Rand Index on Training Set: 0.2721869804033402\n",
      "Adjusted Rand Index on Test Set: 0.25023291288924226\n",
      "Dataset  9\n",
      "ChlorineConcentration\n",
      "Adjusted Rand Index on Training Set: 0.0023870678451229856\n",
      "Adjusted Rand Index on Test Set: -0.005570128709703299\n",
      "Dataset  10\n",
      "CinC_ECG_torso\n",
      "Adjusted Rand Index on Training Set: 0.21052441502984057\n",
      "Adjusted Rand Index on Test Set: 0.23387652109021426\n",
      "Dataset  11\n",
      "Coffee\n",
      "Adjusted Rand Index on Training Set: 0.49118866073970957\n",
      "Adjusted Rand Index on Test Set: 0.5878962536023055\n",
      "Dataset  12\n",
      "Computers\n",
      "Adjusted Rand Index on Training Set: 0.004051245885803447\n",
      "Adjusted Rand Index on Test Set: -0.0048484848484848485\n",
      "Dataset  13\n",
      "Cricket_X\n",
      "Adjusted Rand Index on Training Set: 0.07313654823204094\n",
      "Adjusted Rand Index on Test Set: 0.061293323068682085\n",
      "Dataset  14\n",
      "Cricket_Y\n",
      "Adjusted Rand Index on Training Set: 0.09023990690504419\n",
      "Adjusted Rand Index on Test Set: 0.09118374824208038\n",
      "Dataset  15\n",
      "Cricket_Z\n",
      "Adjusted Rand Index on Training Set: 0.056647311636158514\n",
      "Adjusted Rand Index on Test Set: 0.036956278087196905\n",
      "Dataset  16\n",
      "DiatomSizeReduction\n",
      "Adjusted Rand Index on Training Set: 0.8189244572859417\n",
      "Adjusted Rand Index on Test Set: 0.8796741947426879\n",
      "Dataset  17\n",
      "DistalPhalanxOutlineAgeGroup\n",
      "Adjusted Rand Index on Training Set: 0.26282183532854003\n",
      "Adjusted Rand Index on Test Set: 0.22708010226779612\n",
      "Dataset  18\n",
      "DistalPhalanxOutlineCorrect\n",
      "Adjusted Rand Index on Training Set: 0.03869488397176492\n",
      "Adjusted Rand Index on Test Set: 0.0575538429226134\n",
      "Dataset  19\n",
      "DistalPhalanxTW\n",
      "Adjusted Rand Index on Training Set: 0.4243171837449507\n",
      "Adjusted Rand Index on Test Set: 0.3809567397792045\n",
      "Dataset  20\n",
      "Earthquakes\n",
      "Adjusted Rand Index on Training Set: -0.0019431574532234586\n",
      "Adjusted Rand Index on Test Set: 0.006573484664534049\n",
      "Dataset  21\n",
      "ECG200\n",
      "Adjusted Rand Index on Training Set: 0.4100434351901242\n",
      "Adjusted Rand Index on Test Set: 0.2373712604217754\n",
      "Dataset  22\n",
      "ECG5000\n",
      "Adjusted Rand Index on Training Set: 0.42038397904074565\n",
      "Adjusted Rand Index on Test Set: 0.42769378389494733\n",
      "Dataset  23\n",
      "ECGFiveDays\n",
      "Adjusted Rand Index on Training Set: 0.003670239528847993\n",
      "Adjusted Rand Index on Test Set: -0.004520907838344356\n",
      "Dataset  24\n",
      "ElectricDevices\n",
      "Adjusted Rand Index on Training Set: 0.16227808140617175\n",
      "Adjusted Rand Index on Test Set: 0.1648865124154441\n",
      "Dataset  25\n",
      "FaceAll\n",
      "Adjusted Rand Index on Training Set: 0.11297650378263607\n",
      "Adjusted Rand Index on Test Set: 0.095250210051182\n",
      "Dataset  26\n",
      "FaceFour\n",
      "Adjusted Rand Index on Training Set: 0.30439715877912327\n",
      "Adjusted Rand Index on Test Set: 0.3888620492775265\n",
      "Dataset  27\n",
      "FacesUCR\n",
      "Adjusted Rand Index on Training Set: 0.10019642097712311\n",
      "Adjusted Rand Index on Test Set: 0.13491121452889757\n",
      "Dataset  28\n",
      "FISH\n",
      "Adjusted Rand Index on Training Set: 0.15502169885020875\n",
      "Adjusted Rand Index on Test Set: 0.04832940263323387\n",
      "Dataset  29\n",
      "FordA\n",
      "Adjusted Rand Index on Training Set: 0.010129082664755259\n",
      "Adjusted Rand Index on Test Set: 0.009697440285065511\n",
      "Dataset  30\n",
      "FordB\n",
      "Adjusted Rand Index on Training Set: 0.00033380003698735495\n",
      "Adjusted Rand Index on Test Set: 5.265263499026267e-05\n",
      "Dataset  31\n",
      "Gun_Point\n",
      "Adjusted Rand Index on Training Set: 0.09386470740587478\n",
      "Adjusted Rand Index on Test Set: 0.07111882046834347\n",
      "Dataset  32\n",
      "Ham\n",
      "Adjusted Rand Index on Training Set: 0.03508437266757478\n",
      "Adjusted Rand Index on Test Set: 0.007104795737122539\n",
      "Dataset  33\n",
      "HandOutlines\n",
      "Adjusted Rand Index on Training Set: 0.20585079385090616\n",
      "Adjusted Rand Index on Test Set: 0.16010422635928878\n",
      "Dataset  34\n",
      "Haptics\n",
      "Adjusted Rand Index on Training Set: 0.046274123431724094\n",
      "Adjusted Rand Index on Test Set: 0.08090564850130505\n",
      "Dataset  35\n",
      "Herring\n",
      "Adjusted Rand Index on Training Set: 0.0033923029053770496\n",
      "Adjusted Rand Index on Test Set: -0.0317826295159889\n",
      "Dataset  36\n",
      "InlineSkate\n",
      "Adjusted Rand Index on Training Set: 0.01117734227995157\n",
      "Adjusted Rand Index on Test Set: 0.009951768839615037\n",
      "Dataset  37\n",
      "InsectWingbeatSound\n",
      "Adjusted Rand Index on Training Set: 0.32268832409524595\n",
      "Adjusted Rand Index on Test Set: 0.3126926861096694\n",
      "Dataset  38\n",
      "ItalyPowerDemand\n",
      "Adjusted Rand Index on Training Set: 0.5165187047586893\n",
      "Adjusted Rand Index on Test Set: 0.5472545942561559\n",
      "Dataset  39\n",
      "LargeKitchenAppliances\n",
      "Adjusted Rand Index on Training Set: 0.022006383493282933\n",
      "Adjusted Rand Index on Test Set: 0.001392483289899483\n",
      "Dataset  40\n",
      "Lighting2\n",
      "Adjusted Rand Index on Training Set: 0.057948290963353806\n",
      "Adjusted Rand Index on Test Set: 0.10830512887268942\n",
      "Dataset  41\n",
      "Lighting7\n",
      "Adjusted Rand Index on Training Set: 0.19225217328142838\n",
      "Adjusted Rand Index on Test Set: 0.29078204783660305\n",
      "Dataset  42\n",
      "MALLAT\n",
      "Adjusted Rand Index on Training Set: 0.5955949415438153\n",
      "Adjusted Rand Index on Test Set: 0.6034461544951523\n",
      "Dataset  43\n",
      "Meat\n",
      "Adjusted Rand Index on Training Set: 0.6269155567439295\n",
      "Adjusted Rand Index on Test Set: 0.5750969887308331\n",
      "Dataset  44\n",
      "MedicalImages\n",
      "Adjusted Rand Index on Training Set: 0.035097208261349995\n",
      "Adjusted Rand Index on Test Set: 0.04512947961930594\n",
      "Dataset  45\n",
      "MiddlePhalanxOutlineAgeGroup\n",
      "Adjusted Rand Index on Training Set: 0.4420152994673162\n",
      "Adjusted Rand Index on Test Set: 0.3963962522397071\n",
      "Dataset  46\n",
      "MiddlePhalanxOutlineCorrect\n",
      "Adjusted Rand Index on Training Set: 0.056194556848172594\n",
      "Adjusted Rand Index on Test Set: 0.04885303694357658\n",
      "Dataset  47\n",
      "MiddlePhalanxTW\n",
      "Adjusted Rand Index on Training Set: 0.34302185438060384\n",
      "Adjusted Rand Index on Test Set: 0.3054811520534582\n",
      "Dataset  48\n",
      "MoteStrain\n",
      "Adjusted Rand Index on Training Set: 0.2521483651831544\n",
      "Adjusted Rand Index on Test Set: 0.3305706983183208\n",
      "Dataset  49\n",
      "NonInvasiveFatalECG_Thorax1\n",
      "Adjusted Rand Index on Training Set: 0.11000937739741899\n",
      "Adjusted Rand Index on Test Set: 0.10578346902813682\n",
      "Dataset  50\n",
      "NonInvasiveFatalECG_Thorax2\n",
      "Adjusted Rand Index on Training Set: 0.09718825902410402\n",
      "Adjusted Rand Index on Test Set: 0.09542870434320082\n",
      "Dataset  51\n",
      "OliveOil\n",
      "Adjusted Rand Index on Training Set: 0.714742274086671\n",
      "Adjusted Rand Index on Test Set: 0.6168359941944848\n",
      "Dataset  52\n",
      "OSULeaf\n",
      "Adjusted Rand Index on Training Set: 0.16018905182251295\n",
      "Adjusted Rand Index on Test Set: 0.08675255467210909\n",
      "Dataset  53\n",
      "PhalangesOutlinesCorrect\n",
      "Adjusted Rand Index on Training Set: 0.0038726304522774687\n",
      "Adjusted Rand Index on Test Set: 0.0013160838236778846\n",
      "Dataset  54\n",
      "Phoneme\n",
      "Adjusted Rand Index on Training Set: 0.0023477362638542345\n",
      "Adjusted Rand Index on Test Set: 0.0011087688453910733\n",
      "Dataset  55\n",
      "Plane\n",
      "Adjusted Rand Index on Training Set: 0.5243031747991824\n",
      "Adjusted Rand Index on Test Set: 0.46761929999796154\n",
      "Dataset  56\n",
      "ProximalPhalanxOutlineAgeGroup\n",
      "Adjusted Rand Index on Training Set: 0.3419463466006671\n",
      "Adjusted Rand Index on Test Set: 0.27396485914616436\n",
      "Dataset  57\n",
      "ProximalPhalanxOutlineCorrect\n",
      "Adjusted Rand Index on Training Set: 0.04243880863466088\n",
      "Adjusted Rand Index on Test Set: 0.0744241809410339\n",
      "Dataset  58\n",
      "ProximalPhalanxTW\n",
      "Adjusted Rand Index on Training Set: 0.405864982306469\n",
      "Adjusted Rand Index on Test Set: 0.4499796413045999\n",
      "Dataset  59\n",
      "RefrigerationDevices\n",
      "Adjusted Rand Index on Training Set: -0.0011304648768439938\n",
      "Adjusted Rand Index on Test Set: -0.0009582164444243453\n",
      "Dataset  60\n",
      "ScreenType\n",
      "Adjusted Rand Index on Training Set: 0.012521915807585025\n",
      "Adjusted Rand Index on Test Set: 0.009001186522740318\n",
      "Dataset  61\n",
      "ShapeletSim\n",
      "Adjusted Rand Index on Training Set: -0.0030513667343154546\n",
      "Adjusted Rand Index on Test Set: 0.0019811788013868533\n",
      "Dataset  62\n",
      "ShapesAll\n",
      "Adjusted Rand Index on Training Set: 0.05231679633940989\n",
      "Adjusted Rand Index on Test Set: 0.04146063996954582\n",
      "Dataset  63\n",
      "SmallKitchenAppliances\n",
      "Adjusted Rand Index on Training Set: 0.0008930092429296139\n",
      "Adjusted Rand Index on Test Set: 0.004989916398183328\n",
      "Dataset  64\n",
      "SonyAIBORobotSurface\n",
      "Adjusted Rand Index on Training Set: 0.24933298488666072\n",
      "Adjusted Rand Index on Test Set: 0.32535294772149903\n",
      "Dataset  65\n",
      "SonyAIBORobotSurfaceII\n",
      "Adjusted Rand Index on Training Set: 0.29260432896668265\n",
      "Adjusted Rand Index on Test Set: 0.21019256694638105\n",
      "Dataset  66\n",
      "StarLightCurves\n",
      "Adjusted Rand Index on Training Set: 0.3682776811025087\n",
      "Adjusted Rand Index on Test Set: 0.34878036775665355\n",
      "Dataset  67\n",
      "Strawberry\n",
      "Adjusted Rand Index on Training Set: 0.008839702094482175\n",
      "Adjusted Rand Index on Test Set: 0.0016505425902195766\n",
      "Dataset  68\n",
      "SwedishLeaf\n",
      "Adjusted Rand Index on Training Set: 0.14152304219571812\n",
      "Adjusted Rand Index on Test Set: 0.13705236145182997\n",
      "Dataset  69\n",
      "Symbols\n",
      "Adjusted Rand Index on Training Set: 0.6701646127914067\n",
      "Adjusted Rand Index on Test Set: 0.7279834416195335\n",
      "Dataset  70\n",
      "synthetic_control\n",
      "Adjusted Rand Index on Training Set: 0.5750325195940434\n",
      "Adjusted Rand Index on Test Set: 0.5580041604928748\n",
      "Dataset  71\n",
      "ToeSegmentation1\n",
      "Adjusted Rand Index on Training Set: -0.0015050692113936254\n",
      "Adjusted Rand Index on Test Set: 0.02905047652665016\n",
      "Dataset  72\n",
      "ToeSegmentation2\n",
      "Adjusted Rand Index on Training Set: 0.014471208742333866\n",
      "Adjusted Rand Index on Test Set: -0.021098314867752904\n",
      "Dataset  73\n",
      "Trace\n",
      "Adjusted Rand Index on Training Set: 0.2928636611551775\n",
      "Adjusted Rand Index on Test Set: 0.33862433862433855\n",
      "Dataset  74\n",
      "TwoLeadECG\n",
      "Adjusted Rand Index on Training Set: 0.004187478885172455\n",
      "Adjusted Rand Index on Test Set: -0.0010030860132344546\n",
      "Dataset  75\n",
      "Two_Patterns\n",
      "Adjusted Rand Index on Training Set: 0.020085428454392873\n",
      "Adjusted Rand Index on Test Set: 0.026208561833555053\n",
      "Dataset  76\n",
      "UWaveGestureLibraryAll\n",
      "Adjusted Rand Index on Training Set: 0.47791036081616656\n",
      "Adjusted Rand Index on Test Set: 0.46283927294029054\n",
      "Dataset  77\n",
      "uWaveGestureLibrary_X\n",
      "Adjusted Rand Index on Training Set: 0.3299648595114634\n",
      "Adjusted Rand Index on Test Set: 0.33622522605612354\n",
      "Dataset  78\n",
      "uWaveGestureLibrary_Y\n",
      "Adjusted Rand Index on Training Set: 0.33675638875880454\n",
      "Adjusted Rand Index on Test Set: 0.3213004461218034\n",
      "Dataset  79\n",
      "uWaveGestureLibrary_Z\n",
      "Adjusted Rand Index on Training Set: 0.31771068800777097\n",
      "Adjusted Rand Index on Test Set: 0.2946489484279623\n",
      "Dataset  80\n",
      "wafer\n",
      "Adjusted Rand Index on Training Set: 0.05569448738524084\n",
      "Adjusted Rand Index on Test Set: 0.058766859149983225\n",
      "Dataset  81\n",
      "Wine\n",
      "Adjusted Rand Index on Training Set: 0.0012341182303352719\n",
      "Adjusted Rand Index on Test Set: -0.030090972708187547\n",
      "Dataset  82\n",
      "WordsSynonyms\n",
      "Adjusted Rand Index on Training Set: 0.10947462940710878\n",
      "Adjusted Rand Index on Test Set: 0.10422118835287301\n",
      "Dataset  83\n",
      "Worms\n",
      "Adjusted Rand Index on Training Set: -0.002845565377816157\n",
      "Adjusted Rand Index on Test Set: 0.0182312308911788\n",
      "Dataset  84\n",
      "WormsTwoClass\n",
      "Adjusted Rand Index on Training Set: 0.00440069789550413\n",
      "Adjusted Rand Index on Test Set: 0.016764655320612016\n",
      "Dataset  85\n",
      "yoga\n",
      "Adjusted Rand Index on Training Set: 0.004227036733026657\n",
      "Adjusted Rand Index on Test Set: 0.004192782410311586\n",
      "Elapsed: 2.9342538197835286 min \n"
     ]
    }
   ],
   "source": [
    "# Train and Evaluate k-Means\n",
    "timer = ElapsedTimer()\n",
    "cnt = 0\n",
    "for i in d:\n",
    "    cnt += 1\n",
    "    print(\"Dataset \", cnt)\n",
    "    newpath = os.path.sep.join([mypath, i, \"\"])\n",
    "    onlyfiles = [f for f in listdir(newpath) if isfile(join(newpath, f))]\n",
    "    j = onlyfiles[0]\n",
    "    k = onlyfiles[1]\n",
    "    data_train = np.loadtxt(newpath+j, delimiter=\",\")\n",
    "    data_test = np.loadtxt(newpath+k, delimiter=\",\")\n",
    "\n",
    "    data_joined = np.concatenate((data_train,data_test),axis=0)\n",
    "    data_train, data_test = train_test_split(data_joined, \\\n",
    "                                        test_size=0.20, random_state=2019)\n",
    "\n",
    "    X_train = to_time_series_dataset(data_train[:, 1:])\n",
    "    y_train = data_train[:, 0].astype(np.int)\n",
    "    X_test = to_time_series_dataset(data_test[:, 1:])\n",
    "    y_test = data_test[:, 0].astype(np.int)\n",
    "    \n",
    "    X_train = TimeSeriesScalerMeanVariance(mu=0., std=1.) \\\n",
    "                                    .fit_transform(X_train)\n",
    "    X_test = TimeSeriesScalerMeanVariance(mu=0., std=1.) \\\n",
    "                                    .fit_transform(X_test)\n",
    "    \n",
    "    classes = len(np.unique(data_train[:,0]))\n",
    "    km = TimeSeriesKMeans(n_clusters=5, max_iter=10, n_init=10, \\\n",
    "                          metric=\"euclidean\", verbose=0, random_state=2019)\n",
    "    km.fit(X_train)\n",
    "\n",
    "    print(i)\n",
    "    preds = km.predict(X_train)\n",
    "    ars = adjusted_rand_score(data_train[:,0],preds)\n",
    "    print(\"Adjusted Rand Index on Training Set:\", ars)\n",
    "    kMeansDF.loc[i,\"Train ARS\"] = ars\n",
    "    \n",
    "    preds_test = km.predict(X_test)\n",
    "    ars = adjusted_rand_score(data_test[:,0],preds_test)\n",
    "    print(\"Adjusted Rand Index on Test Set:\", ars)\n",
    "    kMeansDF.loc[i,\"Test ARS\"] = ars\n",
    "    \n",
    "kMeansTime = timer.elapsed_time()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 77,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Time to Run k-Means Experiment in Minutes: 2.9342538197835286\n"
     ]
    }
   ],
   "source": [
    "print(\"Time to Run k-Means Experiment in Minutes:\", kMeansTime/60)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 78,
   "metadata": {
    "collapsed": true,
    "jupyter": {
     "outputs_hidden": true
    }
   },
   "outputs": [],
   "source": [
    "kMeansDF.to_pickle(os.path.sep.join([current_path, \"datasets\", \"pickled_data\", \\\n",
    "                                     \"kMeansDF.pickle\"]))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 79,
   "metadata": {
    "collapsed": true,
    "jupyter": {
     "outputs_hidden": true
    }
   },
   "outputs": [],
   "source": [
    "# HDBSCAN Experiment - FULL RUN"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 80,
   "metadata": {
    "collapsed": true,
    "jupyter": {
     "outputs_hidden": true
    }
   },
   "outputs": [],
   "source": [
    "# Create dataframe\n",
    "hdbscanDF = pd.DataFrame(data=[],index=[v for v in d], \\\n",
    "                         columns=[\"Train ARS\",\"Test ARS\"])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 81,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Dataset  1\n",
      "50words\n",
      "Adjusted Rand Index on Training Set: 0.04712217317560057\n",
      "Adjusted Rand Index on Test Set: 0.01919753314304532\n",
      "Dataset  2\n",
      "Adiac\n",
      "Adjusted Rand Index on Training Set: 0.007247077169102767\n",
      "Adjusted Rand Index on Test Set: 0.005341355551188271\n",
      "Dataset  3\n",
      "ArrowHead\n",
      "Adjusted Rand Index on Training Set: 0.08680758632720399\n",
      "Adjusted Rand Index on Test Set: 0.054936310643678835\n",
      "Dataset  4\n",
      "Beef\n",
      "Adjusted Rand Index on Training Set: 0.07675278820826091\n",
      "Adjusted Rand Index on Test Set: 0.094885100074129\n",
      "Dataset  5\n",
      "BeetleFly\n",
      "Adjusted Rand Index on Training Set: 0.0\n",
      "Adjusted Rand Index on Test Set: 0.0\n",
      "Dataset  6\n",
      "BirdChicken\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\ankur_\\Anaconda3\\envs\\unsupervisedLearning\\lib\\site-packages\\hdbscan\\prediction.py:385: UserWarning: Clusterer does not have any defined clusters, new data will be automatically predicted as noise.\n",
      "  warn('Clusterer does not have any defined clusters, new data'\n",
      "C:\\Users\\ankur_\\Anaconda3\\envs\\unsupervisedLearning\\lib\\site-packages\\hdbscan\\prediction.py:385: UserWarning: Clusterer does not have any defined clusters, new data will be automatically predicted as noise.\n",
      "  warn('Clusterer does not have any defined clusters, new data'\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Adjusted Rand Index on Training Set: 0.0\n",
      "Adjusted Rand Index on Test Set: 0.0\n",
      "Dataset  7\n",
      "Car\n",
      "Adjusted Rand Index on Training Set: 0.1327460764307008\n",
      "Adjusted Rand Index on Test Set: 0.053611973495204156\n",
      "Dataset  8\n",
      "CBF\n",
      "Adjusted Rand Index on Training Set: 0.23824326045039873\n",
      "Adjusted Rand Index on Test Set: 0.24127819753652197\n",
      "Dataset  9\n",
      "ChlorineConcentration\n",
      "Adjusted Rand Index on Training Set: 0.03471687057268341\n",
      "Adjusted Rand Index on Test Set: 0.03360750032547304\n",
      "Dataset  10\n",
      "CinC_ECG_torso\n",
      "Adjusted Rand Index on Training Set: 0.5942196748524765\n",
      "Adjusted Rand Index on Test Set: 0.5958171238376715\n",
      "Dataset  11\n",
      "Coffee\n",
      "Adjusted Rand Index on Training Set: 0.6425515836066016\n",
      "Adjusted Rand Index on Test Set: 0.5272206303724928\n",
      "Dataset  12\n",
      "Computers\n",
      "Adjusted Rand Index on Training Set: -0.0013452895579769142\n",
      "Adjusted Rand Index on Test Set: -0.006868686868686869\n",
      "Dataset  13\n",
      "Cricket_X\n",
      "Adjusted Rand Index on Training Set: 0.02577712650416776\n",
      "Adjusted Rand Index on Test Set: 0.042728088683931666\n",
      "Dataset  14\n",
      "Cricket_Y\n",
      "Adjusted Rand Index on Training Set: 0.029168876587940432\n",
      "Adjusted Rand Index on Test Set: 0.008648947220127375\n",
      "Dataset  15\n",
      "Cricket_Z\n",
      "Adjusted Rand Index on Training Set: 0.03972087339342956\n",
      "Adjusted Rand Index on Test Set: 0.048097894508440785\n",
      "Dataset  16\n",
      "DiatomSizeReduction\n",
      "Adjusted Rand Index on Training Set: 0.9071439403134605\n",
      "Adjusted Rand Index on Test Set: 0.9644709995712017\n",
      "Dataset  17\n",
      "DistalPhalanxOutlineAgeGroup\n",
      "Adjusted Rand Index on Training Set: 0.44849852886593217\n",
      "Adjusted Rand Index on Test Set: 0.3667389470954875\n",
      "Dataset  18\n",
      "DistalPhalanxOutlineCorrect\n",
      "Adjusted Rand Index on Training Set: 0.09393532492277866\n",
      "Adjusted Rand Index on Test Set: 0.06354413349309822\n",
      "Dataset  19\n",
      "DistalPhalanxTW\n",
      "Adjusted Rand Index on Training Set: 0.6534997242309845\n",
      "Adjusted Rand Index on Test Set: 0.5698434654306719\n",
      "Dataset  20\n",
      "Earthquakes\n",
      "Adjusted Rand Index on Training Set: 0.010928035357605876\n",
      "Adjusted Rand Index on Test Set: 0.0069634933425928676\n",
      "Dataset  21\n",
      "ECG200\n",
      "Adjusted Rand Index on Training Set: 0.41995430312261994\n",
      "Adjusted Rand Index on Test Set: 0.2675080425637219\n",
      "Dataset  22\n",
      "ECG5000\n",
      "Adjusted Rand Index on Training Set: 0.7556104485290581\n",
      "Adjusted Rand Index on Test Set: 0.7561234640907506\n",
      "Dataset  23\n",
      "ECGFiveDays\n",
      "Adjusted Rand Index on Training Set: 0.16072052207840315\n",
      "Adjusted Rand Index on Test Set: 0.17272704427554567\n",
      "Dataset  24\n",
      "ElectricDevices\n",
      "Adjusted Rand Index on Training Set: 0.003080324382840311\n",
      "Adjusted Rand Index on Test Set: 0.004862356920630733\n",
      "Dataset  25\n",
      "FaceAll\n",
      "Adjusted Rand Index on Training Set: 0.10711222733884268\n",
      "Adjusted Rand Index on Test Set: 0.07375216893457956\n",
      "Dataset  26\n",
      "FaceFour\n",
      "Adjusted Rand Index on Training Set: 0.17037443595419818\n",
      "Adjusted Rand Index on Test Set: 0.13958760991451943\n",
      "Dataset  27\n",
      "FacesUCR\n",
      "Adjusted Rand Index on Training Set: 0.10797422103102844\n",
      "Adjusted Rand Index on Test Set: 0.08636985958018598\n",
      "Dataset  28\n",
      "FISH\n",
      "Adjusted Rand Index on Training Set: 0.1604256783126502\n",
      "Adjusted Rand Index on Test Set: 0.02498014230202374\n",
      "Dataset  29\n",
      "FordA\n",
      "Adjusted Rand Index on Training Set: 0.004283193224377136\n",
      "Adjusted Rand Index on Test Set: 0.00023211062236586075\n",
      "Dataset  30\n",
      "FordB\n",
      "Adjusted Rand Index on Training Set: -0.00074209928219292\n",
      "Adjusted Rand Index on Test Set: 0.00011947224748877147\n",
      "Dataset  31\n",
      "Gun_Point\n",
      "Adjusted Rand Index on Training Set: 0.17534633640824643\n",
      "Adjusted Rand Index on Test Set: 0.19900299102691926\n",
      "Dataset  32\n",
      "Ham\n",
      "Adjusted Rand Index on Training Set: 0.0\n",
      "Adjusted Rand Index on Test Set: 0.0\n",
      "Dataset  33\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\ankur_\\Anaconda3\\envs\\unsupervisedLearning\\lib\\site-packages\\hdbscan\\prediction.py:385: UserWarning: Clusterer does not have any defined clusters, new data will be automatically predicted as noise.\n",
      "  warn('Clusterer does not have any defined clusters, new data'\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "HandOutlines\n",
      "Adjusted Rand Index on Training Set: 0.029797481009491162\n",
      "Adjusted Rand Index on Test Set: 0.019686617807504105\n",
      "Dataset  34\n",
      "Haptics\n",
      "Adjusted Rand Index on Training Set: 0.02733461700411107\n",
      "Adjusted Rand Index on Test Set: 0.06337061391206895\n",
      "Dataset  35\n",
      "Herring\n",
      "Adjusted Rand Index on Training Set: 0.0\n",
      "Adjusted Rand Index on Test Set: 0.0\n",
      "Dataset  36\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\ankur_\\Anaconda3\\envs\\unsupervisedLearning\\lib\\site-packages\\hdbscan\\prediction.py:385: UserWarning: Clusterer does not have any defined clusters, new data will be automatically predicted as noise.\n",
      "  warn('Clusterer does not have any defined clusters, new data'\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "InlineSkate\n",
      "Adjusted Rand Index on Training Set: 0.007028878666084085\n",
      "Adjusted Rand Index on Test Set: 0.018667966408012385\n",
      "Dataset  37\n",
      "InsectWingbeatSound\n",
      "Adjusted Rand Index on Training Set: 0.0968567695614674\n",
      "Adjusted Rand Index on Test Set: 0.0536686991371905\n",
      "Dataset  38\n",
      "ItalyPowerDemand\n",
      "Adjusted Rand Index on Training Set: 0.21466704676509\n",
      "Adjusted Rand Index on Test Set: 0.20486940656857858\n",
      "Dataset  39\n",
      "LargeKitchenAppliances\n",
      "Adjusted Rand Index on Training Set: 0.013751576907123196\n",
      "Adjusted Rand Index on Test Set: 0.008171148245798801\n",
      "Dataset  40\n",
      "Lighting2\n",
      "Adjusted Rand Index on Training Set: 0.05613360146891671\n",
      "Adjusted Rand Index on Test Set: -0.03833865814696479\n",
      "Dataset  41\n",
      "Lighting7\n",
      "Adjusted Rand Index on Training Set: 0.14953599314276508\n",
      "Adjusted Rand Index on Test Set: 0.16885359116022097\n",
      "Dataset  42\n",
      "MALLAT\n",
      "Adjusted Rand Index on Training Set: 0.6188384769726322\n",
      "Adjusted Rand Index on Test Set: 0.5537161824689919\n",
      "Dataset  43\n",
      "Meat\n",
      "Adjusted Rand Index on Training Set: 0.3476487772130403\n",
      "Adjusted Rand Index on Test Set: 0.3015349067502889\n",
      "Dataset  44\n",
      "MedicalImages\n",
      "Adjusted Rand Index on Training Set: -0.150038201550172\n",
      "Adjusted Rand Index on Test Set: -0.11785577580748442\n",
      "Dataset  45\n",
      "MiddlePhalanxOutlineAgeGroup\n",
      "Adjusted Rand Index on Training Set: 0.41765554916607006\n",
      "Adjusted Rand Index on Test Set: 0.39497787362928377\n",
      "Dataset  46\n",
      "MiddlePhalanxOutlineCorrect\n",
      "Adjusted Rand Index on Training Set: 0.015681734627540963\n",
      "Adjusted Rand Index on Test Set: 0.011636393161816829\n",
      "Dataset  47\n",
      "MiddlePhalanxTW\n",
      "Adjusted Rand Index on Training Set: 0.43682311005267477\n",
      "Adjusted Rand Index on Test Set: 0.41563266126864085\n",
      "Dataset  48\n",
      "MoteStrain\n",
      "Adjusted Rand Index on Training Set: 0.023523425907611924\n",
      "Adjusted Rand Index on Test Set: 0.00913196670794574\n",
      "Dataset  49\n",
      "NonInvasiveFatalECG_Thorax1\n",
      "Adjusted Rand Index on Training Set: 0.07116679511792358\n",
      "Adjusted Rand Index on Test Set: 0.06753522800746767\n",
      "Dataset  50\n",
      "NonInvasiveFatalECG_Thorax2\n",
      "Adjusted Rand Index on Training Set: 0.23704799176948152\n",
      "Adjusted Rand Index on Test Set: 0.20164086261042938\n",
      "Dataset  51\n",
      "OliveOil\n",
      "Adjusted Rand Index on Training Set: 0.5495176515334877\n",
      "Adjusted Rand Index on Test Set: 0.1448763250883392\n",
      "Dataset  52\n",
      "OSULeaf\n",
      "Adjusted Rand Index on Training Set: 0.04001478938901476\n",
      "Adjusted Rand Index on Test Set: 0.02723300352786901\n",
      "Dataset  53\n",
      "PhalangesOutlinesCorrect\n",
      "Adjusted Rand Index on Training Set: 0.005551200182238192\n",
      "Adjusted Rand Index on Test Set: 0.011936637554826941\n",
      "Dataset  54\n",
      "Phoneme\n",
      "Adjusted Rand Index on Training Set: 0.0073787800650459414\n",
      "Adjusted Rand Index on Test Set: 0.006566629090881971\n",
      "Dataset  55\n",
      "Plane\n",
      "Adjusted Rand Index on Training Set: 0.8220409725918153\n",
      "Adjusted Rand Index on Test Set: 0.8338607594936709\n",
      "Dataset  56\n",
      "ProximalPhalanxOutlineAgeGroup\n",
      "Adjusted Rand Index on Training Set: 0.5525206412261244\n",
      "Adjusted Rand Index on Test Set: 0.49069856544677004\n",
      "Dataset  57\n",
      "ProximalPhalanxOutlineCorrect\n",
      "Adjusted Rand Index on Training Set: 0.01767326368681402\n",
      "Adjusted Rand Index on Test Set: 0.05682523773427964\n",
      "Dataset  58\n",
      "ProximalPhalanxTW\n",
      "Adjusted Rand Index on Training Set: 0.5629387103192184\n",
      "Adjusted Rand Index on Test Set: 0.5745613245803115\n",
      "Dataset  59\n",
      "RefrigerationDevices\n",
      "Adjusted Rand Index on Training Set: 0.001530734090218992\n",
      "Adjusted Rand Index on Test Set: 0.004970605264586349\n",
      "Dataset  60\n",
      "ScreenType\n",
      "Adjusted Rand Index on Training Set: 0.011413727586245542\n",
      "Adjusted Rand Index on Test Set: 0.014201243097304021\n",
      "Dataset  61\n",
      "ShapeletSim\n",
      "Adjusted Rand Index on Training Set: -0.004611646238182005\n",
      "Adjusted Rand Index on Test Set: 0.0\n",
      "Dataset  62\n",
      "ShapesAll\n",
      "Adjusted Rand Index on Training Set: 0.05355914850022222\n",
      "Adjusted Rand Index on Test Set: 0.04112578753877586\n",
      "Dataset  63\n",
      "SmallKitchenAppliances\n",
      "Adjusted Rand Index on Training Set: 0.009721965984657998\n",
      "Adjusted Rand Index on Test Set: 0.02364354241926606\n",
      "Dataset  64\n",
      "SonyAIBORobotSurface\n",
      "Adjusted Rand Index on Training Set: 0.466820018068988\n",
      "Adjusted Rand Index on Test Set: 0.42266703376456677\n",
      "Dataset  65\n",
      "SonyAIBORobotSurfaceII\n",
      "Adjusted Rand Index on Training Set: 0.4161340005017514\n",
      "Adjusted Rand Index on Test Set: 0.3660889446813046\n",
      "Dataset  66\n",
      "StarLightCurves\n",
      "Adjusted Rand Index on Training Set: 0.6043609835986175\n",
      "Adjusted Rand Index on Test Set: 0.5877318883735247\n",
      "Dataset  67\n",
      "Strawberry\n",
      "Adjusted Rand Index on Training Set: -0.03322716392197656\n",
      "Adjusted Rand Index on Test Set: -0.0534154245785985\n",
      "Dataset  68\n",
      "SwedishLeaf\n",
      "Adjusted Rand Index on Training Set: 0.06172433349002788\n",
      "Adjusted Rand Index on Test Set: 0.02933954935424017\n",
      "Dataset  69\n",
      "Symbols\n",
      "Adjusted Rand Index on Training Set: 0.7109545364381192\n",
      "Adjusted Rand Index on Test Set: 0.7291816499721819\n",
      "Dataset  70\n",
      "synthetic_control\n",
      "Adjusted Rand Index on Training Set: 0.6571978927585601\n",
      "Adjusted Rand Index on Test Set: 0.6385919262409767\n",
      "Dataset  71\n",
      "ToeSegmentation1\n",
      "Adjusted Rand Index on Training Set: 0.09365677519451036\n",
      "Adjusted Rand Index on Test Set: -0.013732221677292938\n",
      "Dataset  72\n",
      "ToeSegmentation2\n",
      "Adjusted Rand Index on Training Set: -0.12025489588150488\n",
      "Adjusted Rand Index on Test Set: -0.1408420341613791\n",
      "Dataset  73\n",
      "Trace\n",
      "Adjusted Rand Index on Training Set: 0.4953963339322084\n",
      "Adjusted Rand Index on Test Set: 0.4826344910339311\n",
      "Dataset  74\n",
      "TwoLeadECG\n",
      "Adjusted Rand Index on Training Set: 0.0487611500065211\n",
      "Adjusted Rand Index on Test Set: 0.023504083041584474\n",
      "Dataset  75\n",
      "Two_Patterns\n",
      "Adjusted Rand Index on Training Set: 0.00063356058457108\n",
      "Adjusted Rand Index on Test Set: 0.0014243955328180008\n",
      "Dataset  76\n",
      "UWaveGestureLibraryAll\n",
      "Adjusted Rand Index on Training Set: 0.06190214181417517\n",
      "Adjusted Rand Index on Test Set: 0.05979770969060198\n",
      "Dataset  77\n",
      "uWaveGestureLibrary_X\n",
      "Adjusted Rand Index on Training Set: 0.02653159583979862\n",
      "Adjusted Rand Index on Test Set: 0.03871896821214037\n",
      "Dataset  78\n",
      "uWaveGestureLibrary_Y\n",
      "Adjusted Rand Index on Training Set: 0.0477090649088894\n",
      "Adjusted Rand Index on Test Set: 0.0421662290798025\n",
      "Dataset  79\n",
      "uWaveGestureLibrary_Z\n",
      "Adjusted Rand Index on Training Set: 0.1413590342301642\n",
      "Adjusted Rand Index on Test Set: 0.10947830831693975\n",
      "Dataset  80\n",
      "wafer\n",
      "Adjusted Rand Index on Training Set: 0.008853695526890217\n",
      "Adjusted Rand Index on Test Set: 0.0304939268405748\n",
      "Dataset  81\n",
      "Wine\n",
      "Adjusted Rand Index on Training Set: -0.003866822564008358\n",
      "Adjusted Rand Index on Test Set: -0.009678534393363306\n",
      "Dataset  82\n",
      "WordsSynonyms\n",
      "Adjusted Rand Index on Training Set: 0.08392638087932967\n",
      "Adjusted Rand Index on Test Set: 0.05427600069796473\n",
      "Dataset  83\n",
      "Worms\n",
      "Adjusted Rand Index on Training Set: 0.06311257887340958\n",
      "Adjusted Rand Index on Test Set: 0.06719316749414765\n",
      "Dataset  84\n",
      "WormsTwoClass\n",
      "Adjusted Rand Index on Training Set: -0.02408917899471696\n",
      "Adjusted Rand Index on Test Set: -0.011352371640797927\n",
      "Dataset  85\n",
      "yoga\n",
      "Adjusted Rand Index on Training Set: 0.01188760764450979\n",
      "Adjusted Rand Index on Test Set: 0.013225605272761549\n",
      "Elapsed: 7.0786975185076395 min \n"
     ]
    }
   ],
   "source": [
    "# Train and Evaluate HDBSCAN\n",
    "timer = ElapsedTimer()\n",
    "cnt = 0\n",
    "for i in d:\n",
    "    cnt += 1\n",
    "    print(\"Dataset \", cnt)\n",
    "    newpath = os.path.sep.join([mypath, i, \"\"])\n",
    "    onlyfiles = [f for f in listdir(newpath) if isfile(join(newpath, f))]\n",
    "    j = onlyfiles[0]\n",
    "    k = onlyfiles[1]\n",
    "    data_train = np.loadtxt(newpath+j, delimiter=\",\")\n",
    "    data_test = np.loadtxt(newpath+k, delimiter=\",\")\n",
    "\n",
    "    data_joined = np.concatenate((data_train,data_test),axis=0)\n",
    "    data_train, data_test = train_test_split(data_joined, \\\n",
    "                                    test_size=0.20, random_state=2019)\n",
    "\n",
    "    X_train = data_train[:, 1:]\n",
    "    y_train = data_train[:, 0].astype(np.int)\n",
    "    X_test = data_test[:, 1:]\n",
    "    y_test = data_test[:, 0].astype(np.int)\n",
    "    \n",
    "    X_train = TimeSeriesScalerMeanVariance(mu=0., std=1.) \\\n",
    "                                    .fit_transform(X_train)\n",
    "    X_test = TimeSeriesScalerMeanVariance(mu=0., std=1.)  \\\n",
    "                                    .fit_transform(X_test)\n",
    "    \n",
    "    classes = len(np.unique(data_train[:,0]))\n",
    "    min_cluster_size = 5\n",
    "    min_samples = None\n",
    "    alpha = 1.0\n",
    "    cluster_selection_method = 'eom'\n",
    "    prediction_data = True\n",
    "\n",
    "    hdb = hdbscan.HDBSCAN(min_cluster_size=min_cluster_size, \\\n",
    "                          min_samples=min_samples, alpha=alpha, \\\n",
    "                          cluster_selection_method= \\\n",
    "                              cluster_selection_method, \\\n",
    "                          prediction_data=prediction_data)\n",
    "\n",
    "    print(i)\n",
    "    preds = hdb.fit_predict(X_train.reshape(X_train.shape[0], \\\n",
    "                                            X_train.shape[1]))\n",
    "    ars = adjusted_rand_score(data_train[:,0],preds)\n",
    "    print(\"Adjusted Rand Index on Training Set:\", ars)\n",
    "    hdbscanDF.loc[i,\"Train ARS\"] = ars\n",
    "    \n",
    "    preds_test = hdbscan.prediction.approximate_predict(hdb, \n",
    "                            X_test.reshape(X_test.shape[0], \\\n",
    "                                           X_test.shape[1]))\n",
    "    ars = adjusted_rand_score(data_test[:,0],preds_test[0])\n",
    "    print(\"Adjusted Rand Index on Test Set:\", ars)\n",
    "    hdbscanDF.loc[i,\"Test ARS\"] = ars\n",
    "    \n",
    "hdbscanTime = timer.elapsed_time()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 82,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Time to Run HDBSCAN Experiment in Minutes: 7.0786975185076395\n"
     ]
    }
   ],
   "source": [
    "print(\"Time to Run HDBSCAN Experiment in Minutes:\", hdbscanTime/60)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 83,
   "metadata": {
    "collapsed": true,
    "jupyter": {
     "outputs_hidden": true
    }
   },
   "outputs": [],
   "source": [
    "hdbscanDF.to_pickle(os.path.sep.join([current_path, \"datasets\", \"pickled_data\", \\\n",
    "                    \"hdbscanDF.pickle\"]))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 84,
   "metadata": {
    "collapsed": true,
    "jupyter": {
     "outputs_hidden": true
    }
   },
   "outputs": [],
   "source": [
    "# Compare All Three Experiments"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 85,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "k-Shape Results\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "Train ARS    0.165139\n",
       "Test ARS     0.151103\n",
       "dtype: float64"
      ]
     },
     "execution_count": 85,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "print(\"k-Shape Results\")\n",
    "kShapeDF.mean()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 86,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "k-Means Results\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "Train ARS    0.184789\n",
       "Test ARS     0.178960\n",
       "dtype: float64"
      ]
     },
     "execution_count": 86,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "print(\"k-Means Results\")\n",
    "kMeansDF.mean()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 87,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "HDBSCAN Results\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "Train ARS    0.178754\n",
       "Test ARS     0.158238\n",
       "dtype: float64"
      ]
     },
     "execution_count": 87,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "print(\"HDBSCAN Results\")\n",
    "hdbscanDF.mean()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 88,
   "metadata": {
    "collapsed": true,
    "jupyter": {
     "outputs_hidden": true
    }
   },
   "outputs": [],
   "source": [
    "# Count top place finishes\n",
    "timeSeriesClusteringDF = pd.DataFrame(data=[],index=kShapeDF.index, \\\n",
    "                            columns=[\"kShapeTest\", \\\n",
    "                                    \"kMeansTest\", \\\n",
    "                                    \"hdbscanTest\"])\n",
    "\n",
    "timeSeriesClusteringDF.kShapeTest = kShapeDF[\"Test ARS\"]\n",
    "timeSeriesClusteringDF.kMeansTest = kMeansDF[\"Test ARS\"]\n",
    "timeSeriesClusteringDF.hdbscanTest = hdbscanDF[\"Test ARS\"]\n",
    "\n",
    "tscResults = timeSeriesClusteringDF.copy()\n",
    "\n",
    "for i in range(0,len(tscResults)):\n",
    "    maxValue = tscResults.iloc[i].max()\n",
    "    tscResults.iloc[i][tscResults.iloc[i]==maxValue]=1\n",
    "    minValue = tscResults .iloc[i].min()\n",
    "    tscResults.iloc[i][tscResults.iloc[i]==minValue]=-1\n",
    "    medianValue = tscResults.iloc[i].median()\n",
    "    tscResults.iloc[i][tscResults.iloc[i]==medianValue]=0"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 89,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>kShape</th>\n",
       "      <th>kMeans</th>\n",
       "      <th>hdbscan</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>firstPlace</th>\n",
       "      <td>31.0</td>\n",
       "      <td>24.0</td>\n",
       "      <td>29.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>secondPlace</th>\n",
       "      <td>19.0</td>\n",
       "      <td>41.0</td>\n",
       "      <td>26.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>thirdPlace</th>\n",
       "      <td>35.0</td>\n",
       "      <td>20.0</td>\n",
       "      <td>30.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "             kShape  kMeans  hdbscan\n",
       "firstPlace     31.0    24.0     29.0\n",
       "secondPlace    19.0    41.0     26.0\n",
       "thirdPlace     35.0    20.0     30.0"
      ]
     },
     "execution_count": 89,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Show results\n",
    "tscResultsDF = pd.DataFrame(data=np.zeros((3,3)), \\\n",
    "                index=[\"firstPlace\",\"secondPlace\",\"thirdPlace\"], \\\n",
    "                columns=[\"kShape\", \"kMeans\",\"hdbscan\"])\n",
    "tscResultsDF.loc[\"firstPlace\",:] = tscResults[tscResults==1].count().values\n",
    "tscResultsDF.loc[\"secondPlace\",:] = tscResults[tscResults==0].count().values\n",
    "tscResultsDF.loc[\"thirdPlace\",:] = tscResults[tscResults==-1].count().values\n",
    "tscResultsDF"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.7.6"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}