Data_Preprocessing.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "8e1bcefc",
   "metadata": {},
   "outputs": [],
   "source": [
    "import random\n",
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "import pandas as pd\n",
    "import os\n",
    "import tensorflow as tf\n",
    "from sklearn.model_selection import train_test_split\n",
    "import matplotlib.pyplot as plt\n",
    "import os\n",
    "import glob\n",
    "import pandas as pd\n",
    "import numpy as np\n",
    "from scipy import interpolate\n",
    "from sklearn.preprocessing import StandardScaler\n",
    "from sklearn.metrics import confusion_matrix\n",
    "from google.colab import drive\n",
    "import pickle\n",
    "from scipy.io import wavfile\n",
    "from scipy.signal import decimate\n",
    "from tensorflow.keras.layers import Input, Conv2D, MaxPooling2D, Flatten, Dense, concatenate, Dropout\n",
    "from tensorflow.keras.models import Model\n",
    "from tensorflow.keras.optimizers import Adam"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "2fde1ee6",
   "metadata": {},
   "outputs": [],
   "source": [
    "### Loading presaved datasets\n",
    "import pickle\n",
    "\n",
    "with open('local_train_x_w.pkl', 'rb') as file:\n",
    "    local_train_x_w = pickle.load(file)\n",
    "with open('local_train_x_p.pkl', 'rb') as file:\n",
    "    local_train_x_p = pickle.load(file)\n",
    "with open('local_train_x_s.pkl', 'rb') as file:\n",
    "    local_train_x_s = pickle.load(file)\n",
    "with open('local_train_y.pkl', 'rb') as file:\n",
    "    local_train_y = pickle.load(file)\n",
    "\n",
    "with open('local_test_x_w.pkl', 'rb') as file:\n",
    "    local_test_x_w = pickle.load(file)\n",
    "with open('local_test_x_p.pkl', 'rb') as file:\n",
    "    local_test_x_p = pickle.load(file)\n",
    "with open('local_test_x_s.pkl', 'rb') as file:\n",
    "    local_test_x_s = pickle.load(file)\n",
    "with open('local_test_y.pkl', 'rb') as file:\n",
    "    local_test_y = pickle.load(file)\n",
    "\n",
    "\n",
    "with open('central_test_x_w.pkl', 'rb') as file:\n",
    "    central_test_x_w = pickle.load(file)\n",
    "with open('central_test_x_p.pkl', 'rb') as file:\n",
    "    central_test_x_p = pickle.load(file)\n",
    "with open('central_test_x_s.pkl', 'rb') as file:\n",
    "    central_test_x_s = pickle.load(file)\n",
    "with open('central_test_y.pkl', 'rb') as file:\n",
    "    central_test_y = pickle.load(file)\n",
    "\n",
    "\n",
    "with open('central_train_x_w.pkl', 'rb') as file:\n",
    "    central_train_x_w = pickle.load(file)\n",
    "with open('central_train_x_p.pkl', 'rb') as file:\n",
    "    central_train_x_p = pickle.load(file)\n",
    "with open('central_train_x_s.pkl', 'rb') as file:\n",
    "    central_train_x_s = pickle.load(file)\n",
    "with open('central_train_y.pkl', 'rb') as file:\n",
    "    central_train_y = pickle.load(file)\n",
    "\n",
    "\n",
    "with open('ood_watch_test.pkl', 'rb') as file:\n",
    "    ood_test_x_w = pickle.load(file)\n",
    "with open('ood_phone_test.pkl', 'rb') as file:\n",
    "    ood_test_x_p = pickle.load(file)\n",
    "with open('ood_audio_test.pkl', 'rb') as file:\n",
    "    ood_test_x_s = pickle.load(file)\n",
    "with open('ood_label_test.pkl', 'rb') as file:\n",
    "    ood_test_y = pickle.load(file)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "bfb67438",
   "metadata": {},
   "source": [
    "# Processing data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "6da2fa0b",
   "metadata": {},
   "outputs": [],
   "source": [
    "for client in range(28):\n",
    "  local_paired_watch[client]=np.reshape(local_paired_watch[client],(-1,100,6,1))\n",
    "  local_paired_phone[client]=np.reshape(local_paired_phone[client],(-1,100,21,1))\n",
    "  local_paired_audio[client]=np.reshape(local_paired_audio[client],(-1,25,32,1))\n",
    "  local_paired_label[client] = np.array(local_paired_label[client])\n",
    "\n",
    "# normalizing local datasets\n",
    "for i in range(28):\n",
    "    shape_w = local_paired_watch[i].shape\n",
    "    shape_p = local_paired_phone[i].shape\n",
    "    shape_s = local_paired_audio[i].shape\n",
    "    x_2d_w = local_paired_watch[i].reshape(shape_w[0], -1)\n",
    "    x_2d_p = local_paired_phone[i].reshape(shape_p[0], -1)\n",
    "    x_2d_s = local_paired_audio[i].reshape(shape_s[0], -1)\n",
    "    # Initialize a scaler\n",
    "    scaler = StandardScaler()\n",
    "    x_normalized_2d = scaler.fit_transform(x_2d_w)\n",
    "    local_paired_watch[i] = x_normalized_2d.reshape(shape_w)\n",
    "\n",
    "    scaler = StandardScaler()\n",
    "    x_normalized_2d = scaler.fit_transform(x_2d_p)\n",
    "    local_paired_phone[i] = x_normalized_2d.reshape(shape_p)\n",
    "\n",
    "    scaler = StandardScaler()\n",
    "    x_normalized_2d = scaler.fit_transform(x_2d_s)\n",
    "    local_paired_audio[i] = x_normalized_2d.reshape(shape_s)\n",
    "\n",
    "del x_2d_p, x_2d_s,x_2d_w, x_normalized_2d"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "7f31005b",
   "metadata": {},
   "outputs": [],
   "source": [
    "##### Data for Federated Learning\n",
    "\n",
    "local_paired_watch_train = list(); local_paired_watch_test = list();\n",
    "local_paired_phone_train = list(); local_paired_phone_test = list();\n",
    "local_paired_audio_train = list(); local_paired_audio_test = list();\n",
    "local_paired_label_train = list(); local_paired_label_test = list();\n",
    "\n",
    "for i in range(20):\n",
    "    x_train_watch, x_test_watch,x_train_phone, x_test_phone,x_train_audio, x_test_audio, y_train, y_test =\\\n",
    "    train_test_split(local_paired_watch[i],local_paired_phone[i],local_paired_audio[i],local_paired_label[i],test_size=0.25,\\\n",
    "                     random_state=42)\n",
    "\n",
    "    local_paired_watch_train.append(x_train_watch)\n",
    "    local_paired_phone_train.append(x_train_phone)\n",
    "    local_paired_audio_train.append(x_train_audio)\n",
    "    local_paired_label_train.append(y_train)\n",
    "\n",
    "    local_paired_watch_test.append(x_test_watch)\n",
    "    local_paired_phone_test.append(x_test_phone)\n",
    "    local_paired_audio_test.append(x_test_audio)\n",
    "    local_paired_label_test.append(y_test)\n",
    "\n",
    "    if i == 0:\n",
    "        global_paired_watch_test = x_test_watch\n",
    "        global_paired_phone_test = x_test_phone\n",
    "        global_paired_audio_test = x_test_audio\n",
    "        global_paired_label_test = y_test\n",
    "    else:\n",
    "        global_paired_watch_test = np.concatenate([global_paired_watch_test,x_test_watch],axis=0)\n",
    "        global_paired_phone_test = np.concatenate([global_paired_phone_test,x_test_phone],axis=0)\n",
    "        global_paired_audio_test = np.concatenate([global_paired_audio_test,x_test_audio],axis=0)\n",
    "        global_paired_label_test = np.concatenate([global_paired_label_test,y_test],axis=0)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "d81e832e",
   "metadata": {},
   "outputs": [],
   "source": [
    "## Data for Centralized Learning\n",
    "\n",
    "for i in range(20):\n",
    "    if i == 0:\n",
    "        central_paired_watch_train = local_paired_watch_train[i]\n",
    "        central_paired_phone_train = local_paired_phone_train[i]\n",
    "        central_paired_audio_train = local_paired_audio_train[i]\n",
    "        central_paired_label_train = local_paired_label_train[i]\n",
    "    else:\n",
    "        central_paired_watch_train = np.concatenate((central_paired_watch_train,local_paired_watch_train[i]),axis=0)\n",
    "        central_paired_phone_train = np.concatenate((central_paired_phone_train,local_paired_phone_train[i]),axis=0)\n",
    "        central_paired_audio_train = np.concatenate((central_paired_audio_train,local_paired_audio_train[i]),axis=0)\n",
    "        central_paired_label_train = np.concatenate((central_paired_label_train,local_paired_label_train[i]),axis=0)\n",
    "\n",
    "for i in range(20,28,1):\n",
    "  if i == 20:\n",
    "    ood_watch_test = local_paired_watch[i]\n",
    "    ood_phone_test = local_paired_phone[i]\n",
    "    ood_audio_test = local_paired_audio[i]\n",
    "    ood_label_test = local_paired_label[i]\n",
    "  else:\n",
    "    ood_watch_test = np.concatenate((ood_watch_test,local_paired_watch[i]),axis=0)\n",
    "    ood_phone_test = np.concatenate((ood_phone_test,local_paired_phone[i]),axis=0)\n",
    "    ood_audio_test = np.concatenate((ood_audio_test,local_paired_audio[i]),axis=0)\n",
    "    ood_label_test = np.concatenate((ood_label_test,local_paired_label[i]),axis=0)"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3 (ipykernel)",
   "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": "3.11.4"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}