USAD.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "etniX_KTlJ5U"
   },
   "source": [
    "# USAD"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "N3jM0qLU8MgZ"
   },
   "source": [
    "## Environment"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "id": "6u1DGKsAlLF-"
   },
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import pandas as pd\n",
    "import matplotlib.pyplot as plt\n",
    "import seaborn as sns\n",
    "import torch\n",
    "import torch.nn as nn\n",
    "\n",
    "from utils import *\n",
    "from usad import *"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 34
    },
    "id": "4AzWlDBI_djV",
    "outputId": "7a8d0c19-2389-461b-c0be-3427a25dda91"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "GPU 0: NVIDIA GeForce GTX 1650 (UUID: GPU-d6079c82-a3ed-1a9d-41f1-cfa122659d29)\n"
     ]
    }
   ],
   "source": [
    "!nvidia-smi -L\n",
    "\n",
    "device = get_default_device()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "1crx5rGP9ONf",
    "jp-MarkdownHeadingCollapsed": true,
    "tags": []
   },
   "source": [
    "## EDA - Data Pre-Processing"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Download dataset"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "kfSj4FYL9W8Y",
    "jp-MarkdownHeadingCollapsed": true,
    "tags": []
   },
   "source": [
    "### Normal period"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/pk/miniconda3/envs/usad/lib/python3.8/site-packages/IPython/core/interactiveshell.py:3444: DtypeWarning: Columns (26) have mixed types.Specify dtype option on import or set low_memory=False.\n",
      "  exec(code_obj, self.user_global_ns, self.user_ns)\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "(495000, 51)"
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#Read data\n",
    "normal = pd.read_csv(\"input/SWaT_Dataset_Normal_v1.csv\")#, nrows=1000)\n",
    "normal = normal.drop([\"Timestamp\" , \"Normal/Attack\" ] , axis = 1)\n",
    "normal.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Transform all columns into float64\n",
    "for i in list(normal): \n",
    "    normal[i]=normal[i].apply(lambda x: str(x).replace(\",\" , \".\"))\n",
    "normal = normal.astype(float)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn import preprocessing\n",
    "min_max_scaler = preprocessing.StandardScaler()\n",
    "\n",
    "x = normal.values\n",
    "x_scaled = min_max_scaler.fit_transform(x)\n",
    "normal = pd.DataFrame(x_scaled)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 87
    },
    "id": "XeDLxV_r1G9n",
    "outputId": "576538dd-64f2-46fa-8e6f-6c2ffdebad15"
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/pk/miniconda3/envs/usad/lib/python3.8/site-packages/IPython/core/interactiveshell.py:3444: DtypeWarning: Columns (26) have mixed types.Specify dtype option on import or set low_memory=False.\n",
      "  exec(code_obj, self.user_global_ns, self.user_ns)\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "(495000, 51)"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#Read data\n",
    "normal = pd.read_csv(\"input/SWaT_Dataset_Normal_v1.csv\")#, nrows=1000)\n",
    "normal = normal.drop([\"Timestamp\" , \"Normal/Attack\" ] , axis = 1)\n",
    "normal.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "id": "fFuLm1GH1G2n"
   },
   "outputs": [],
   "source": [
    "# Transform all columns into float64\n",
    "for i in list(normal): \n",
    "    normal[i]=normal[i].apply(lambda x: str(x).replace(\",\" , \".\"))\n",
    "normal = normal.astype(float)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "zxFNH5kU9hIE"
   },
   "source": [
    "#### Normalization"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "id": "Mfxj4Uxn9kv4"
   },
   "outputs": [],
   "source": [
    "from sklearn import preprocessing\n",
    "min_max_scaler = preprocessing.StandardScaler()\n",
    "\n",
    "x = normal.values\n",
    "x_scaled = min_max_scaler.fit_transform(x)\n",
    "normal = pd.DataFrame(x_scaled)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 126
    },
    "id": "mQ6_U4jn9nlw",
    "outputId": "f1cc1bd6-f1cc-4764-b1cc-2fd989ac4918"
   },
   "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>0</th>\n",
       "      <th>1</th>\n",
       "      <th>2</th>\n",
       "      <th>3</th>\n",
       "      <th>4</th>\n",
       "      <th>5</th>\n",
       "      <th>6</th>\n",
       "      <th>7</th>\n",
       "      <th>8</th>\n",
       "      <th>9</th>\n",
       "      <th>...</th>\n",
       "      <th>41</th>\n",
       "      <th>42</th>\n",
       "      <th>43</th>\n",
       "      <th>44</th>\n",
       "      <th>45</th>\n",
       "      <th>46</th>\n",
       "      <th>47</th>\n",
       "      <th>48</th>\n",
       "      <th>49</th>\n",
       "      <th>50</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>-1.633984</td>\n",
       "      <td>-3.807291</td>\n",
       "      <td>-1.574619</td>\n",
       "      <td>-1.728131</td>\n",
       "      <td>0.0</td>\n",
       "      <td>-2.477683</td>\n",
       "      <td>-0.828035</td>\n",
       "      <td>-0.719672</td>\n",
       "      <td>-1.731442</td>\n",
       "      <td>-1.684036</td>\n",
       "      <td>...</td>\n",
       "      <td>-17.644775</td>\n",
       "      <td>-17.91219</td>\n",
       "      <td>0.0</td>\n",
       "      <td>-17.843131</td>\n",
       "      <td>-5.7965</td>\n",
       "      <td>-17.672513</td>\n",
       "      <td>-0.095148</td>\n",
       "      <td>0.0</td>\n",
       "      <td>-0.089482</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>-1.633984</td>\n",
       "      <td>-3.806646</td>\n",
       "      <td>-1.574619</td>\n",
       "      <td>-1.728131</td>\n",
       "      <td>0.0</td>\n",
       "      <td>-2.477683</td>\n",
       "      <td>-0.828035</td>\n",
       "      <td>-0.719672</td>\n",
       "      <td>-1.731442</td>\n",
       "      <td>-1.684036</td>\n",
       "      <td>...</td>\n",
       "      <td>-17.644775</td>\n",
       "      <td>-17.91219</td>\n",
       "      <td>0.0</td>\n",
       "      <td>-17.843131</td>\n",
       "      <td>-5.7965</td>\n",
       "      <td>-17.672513</td>\n",
       "      <td>-0.095148</td>\n",
       "      <td>0.0</td>\n",
       "      <td>-0.089482</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>2 rows × 51 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "         0         1         2         3    4         5         6         7   \\\n",
       "0 -1.633984 -3.807291 -1.574619 -1.728131  0.0 -2.477683 -0.828035 -0.719672   \n",
       "1 -1.633984 -3.806646 -1.574619 -1.728131  0.0 -2.477683 -0.828035 -0.719672   \n",
       "\n",
       "         8         9   ...         41        42   43         44      45  \\\n",
       "0 -1.731442 -1.684036  ... -17.644775 -17.91219  0.0 -17.843131 -5.7965   \n",
       "1 -1.731442 -1.684036  ... -17.644775 -17.91219  0.0 -17.843131 -5.7965   \n",
       "\n",
       "          46        47   48        49   50  \n",
       "0 -17.672513 -0.095148  0.0 -0.089482  0.0  \n",
       "1 -17.672513 -0.095148  0.0 -0.089482  0.0  \n",
       "\n",
       "[2 rows x 51 columns]"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "normal.head(2)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "_i71RFAi9spa"
   },
   "source": [
    "### Attack"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 87
    },
    "id": "aN_TFp5x9uTE",
    "outputId": "38d7993d-c9a3-461d-c430-ebde697afbc6"
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/pk/miniconda3/envs/usad/lib/python3.8/site-packages/IPython/core/interactiveshell.py:3444: DtypeWarning: Columns (1,9,28,46) have mixed types.Specify dtype option on import or set low_memory=False.\n",
      "  exec(code_obj, self.user_global_ns, self.user_ns)\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "(449919, 51)"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#Read data\n",
    "attack = pd.read_csv(\"input/SWaT_Dataset_Attack_v0.csv\",sep=\";\")#, nrows=1000)\n",
    "labels = [ float(label!= 'Normal' ) for label  in attack[\"Normal/Attack\"].values]\n",
    "attack = attack.drop([\"Timestamp\" , \"Normal/Attack\" ] , axis = 1)\n",
    "attack.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {
    "id": "qLCInT-I9_-D"
   },
   "outputs": [],
   "source": [
    "# Transform all columns into float64\n",
    "for i in list(attack):\n",
    "    attack[i]=attack[i].apply(lambda x: str(x).replace(\",\" , \".\"))\n",
    "attack = attack.astype(float)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "c4cB4v3N-Dhu"
   },
   "source": [
    "#### Normalization"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {
    "id": "jZrha9cO-BGK"
   },
   "outputs": [],
   "source": [
    "from sklearn import preprocessing\n",
    "\n",
    "x = attack.values \n",
    "x_scaled = min_max_scaler.transform(x)\n",
    "attack = pd.DataFrame(x_scaled)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 126
    },
    "id": "z9SwiPco-BUa",
    "outputId": "f2507282-c0f9-4253-ece7-0a802b68240f"
   },
   "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>0</th>\n",
       "      <th>1</th>\n",
       "      <th>2</th>\n",
       "      <th>3</th>\n",
       "      <th>4</th>\n",
       "      <th>5</th>\n",
       "      <th>6</th>\n",
       "      <th>7</th>\n",
       "      <th>8</th>\n",
       "      <th>9</th>\n",
       "      <th>...</th>\n",
       "      <th>41</th>\n",
       "      <th>42</th>\n",
       "      <th>43</th>\n",
       "      <th>44</th>\n",
       "      <th>45</th>\n",
       "      <th>46</th>\n",
       "      <th>47</th>\n",
       "      <th>48</th>\n",
       "      <th>49</th>\n",
       "      <th>50</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0.509078</td>\n",
       "      <td>-0.531668</td>\n",
       "      <td>0.610641</td>\n",
       "      <td>0.57866</td>\n",
       "      <td>0.0</td>\n",
       "      <td>-0.369209</td>\n",
       "      <td>0.09171</td>\n",
       "      <td>-0.399305</td>\n",
       "      <td>0.577082</td>\n",
       "      <td>0.571882</td>\n",
       "      <td>...</td>\n",
       "      <td>0.030545</td>\n",
       "      <td>0.055828</td>\n",
       "      <td>0.0</td>\n",
       "      <td>-0.067305</td>\n",
       "      <td>2.260671</td>\n",
       "      <td>-0.125269</td>\n",
       "      <td>-0.096008</td>\n",
       "      <td>0.0</td>\n",
       "      <td>-0.089482</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0.526046</td>\n",
       "      <td>-0.531345</td>\n",
       "      <td>0.610641</td>\n",
       "      <td>0.57866</td>\n",
       "      <td>0.0</td>\n",
       "      <td>-0.369209</td>\n",
       "      <td>0.09171</td>\n",
       "      <td>-0.399305</td>\n",
       "      <td>0.577082</td>\n",
       "      <td>0.571882</td>\n",
       "      <td>...</td>\n",
       "      <td>0.030545</td>\n",
       "      <td>0.055828</td>\n",
       "      <td>0.0</td>\n",
       "      <td>-0.067305</td>\n",
       "      <td>2.260671</td>\n",
       "      <td>-0.117722</td>\n",
       "      <td>-0.096008</td>\n",
       "      <td>0.0</td>\n",
       "      <td>-0.089482</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>2 rows × 51 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "         0         1         2        3    4         5        6         7   \\\n",
       "0  0.509078 -0.531668  0.610641  0.57866  0.0 -0.369209  0.09171 -0.399305   \n",
       "1  0.526046 -0.531345  0.610641  0.57866  0.0 -0.369209  0.09171 -0.399305   \n",
       "\n",
       "         8         9   ...        41        42   43        44        45  \\\n",
       "0  0.577082  0.571882  ...  0.030545  0.055828  0.0 -0.067305  2.260671   \n",
       "1  0.577082  0.571882  ...  0.030545  0.055828  0.0 -0.067305  2.260671   \n",
       "\n",
       "         46        47   48        49   50  \n",
       "0 -0.125269 -0.096008  0.0 -0.089482  0.0  \n",
       "1 -0.117722 -0.096008  0.0 -0.089482  0.0  \n",
       "\n",
       "[2 rows x 51 columns]"
      ]
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "attack.head(2)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "xXJi503b-j_d"
   },
   "source": [
    "### Windows"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {
    "id": "vyplttZa-BRN"
   },
   "outputs": [],
   "source": [
    "window_size=20"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 34
    },
    "id": "dzGJMp6Y-BN5",
    "outputId": "2949d278-1313-442c-f06b-275a8c6c6578"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(494980, 20, 51)"
      ]
     },
     "execution_count": 15,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "windows_normal=normal.values[np.arange(window_size)[None, :] + np.arange(normal.shape[0]-window_size)[:, None]]\n",
    "windows_normal.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 34
    },
    "id": "17LdB3c8-pRH",
    "outputId": "721059d4-5937-4dd3-d73c-e5d255fc273c"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(449899, 20, 51)"
      ]
     },
     "execution_count": 16,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "windows_attack=attack.values[np.arange(window_size)[None, :] + np.arange(attack.shape[0]-window_size)[:, None]]\n",
    "windows_attack.shape"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "k70ZFxGs-_7m"
   },
   "source": [
    "## Training"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {
    "id": "yi9S0SGnDKNc"
   },
   "outputs": [],
   "source": [
    "import torch.utils.data as data_utils\n",
    "\n",
    "BATCH_SIZE =  1000\n",
    "N_EPOCHS = 40\n",
    "hidden_size = 20\n",
    "\n",
    "w_size=windows_normal.shape[1]*windows_normal.shape[2]\n",
    "z_size= windows_normal.shape[1]*hidden_size\n",
    "\n",
    "windows_normal_train = windows_normal[:int(np.floor(.8 *  windows_normal.shape[0]))]\n",
    "windows_normal_val = windows_normal[int(np.floor(.8 *  windows_normal.shape[0])):int(np.floor(windows_normal.shape[0]))]\n",
    "\n",
    "train_loader = torch.utils.data.DataLoader(data_utils.TensorDataset(\n",
    "    torch.from_numpy(windows_normal_train).float().view(([windows_normal_train.shape[0],w_size]))\n",
    ") , batch_size=BATCH_SIZE, shuffle=False, num_workers=0)\n",
    "\n",
    "val_loader = torch.utils.data.DataLoader(data_utils.TensorDataset(\n",
    "    torch.from_numpy(windows_normal_val).float().view(([windows_normal_val.shape[0],w_size]))\n",
    ") , batch_size=BATCH_SIZE, shuffle=False, num_workers=0)\n",
    "\n",
    "test_loader = torch.utils.data.DataLoader(data_utils.TensorDataset(\n",
    "    torch.from_numpy(windows_attack).float().view(([windows_attack.shape[0],w_size]))\n",
    ") , batch_size=BATCH_SIZE, shuffle=False, num_workers=0)\n",
    "\n",
    "model = UsadModel(w_size, z_size)\n",
    "model = to_device(model,device)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(1020, 400)"
      ]
     },
     "execution_count": 18,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "w_size, z_size"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 1000
    },
    "id": "So9yjDPEDObC",
    "outputId": "629bcd13-37b1-4907-ef0d-46d9e3ad5398",
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch [0], val_loss1: 0.3759, val_loss2: 0.3749\n",
      "Epoch [1], val_loss1: 0.4323, val_loss2: -0.0578\n",
      "Epoch [2], val_loss1: 0.4532, val_loss2: -0.1980\n",
      "Epoch [3], val_loss1: 0.5117, val_loss2: -0.3190\n",
      "Epoch [4], val_loss1: 0.4858, val_loss2: -0.3309\n",
      "Epoch [5], val_loss1: 0.5315, val_loss2: -0.4005\n",
      "Epoch [6], val_loss1: 0.5415, val_loss2: -0.4340\n",
      "Epoch [7], val_loss1: 0.5463, val_loss2: -0.4527\n",
      "Epoch [8], val_loss1: 0.5501, val_loss2: -0.4669\n",
      "Epoch [9], val_loss1: 0.5528, val_loss2: -0.4762\n",
      "Epoch [10], val_loss1: 0.5545, val_loss2: -0.4850\n",
      "Epoch [11], val_loss1: 0.5570, val_loss2: -0.4944\n",
      "Epoch [12], val_loss1: 0.5587, val_loss2: -0.5002\n",
      "Epoch [13], val_loss1: 0.5621, val_loss2: -0.5025\n",
      "Epoch [14], val_loss1: 0.5626, val_loss2: -0.5071\n",
      "Epoch [15], val_loss1: 0.5631, val_loss2: -0.5114\n",
      "Epoch [16], val_loss1: 0.5614, val_loss2: -0.5175\n",
      "Epoch [17], val_loss1: 0.5623, val_loss2: -0.5203\n",
      "Epoch [18], val_loss1: 0.4953, val_loss2: -0.4536\n",
      "Epoch [19], val_loss1: 0.5405, val_loss2: -0.5025\n",
      "Epoch [20], val_loss1: 0.5877, val_loss2: -0.5452\n",
      "Epoch [21], val_loss1: 0.5916, val_loss2: -0.5514\n",
      "Epoch [22], val_loss1: 0.5884, val_loss2: -0.5519\n",
      "Epoch [23], val_loss1: 0.6080, val_loss2: -0.5729\n",
      "Epoch [24], val_loss1: 0.6205, val_loss2: -0.5886\n",
      "Epoch [25], val_loss1: 0.6112, val_loss2: -0.5808\n",
      "Epoch [26], val_loss1: 0.6079, val_loss2: -0.5794\n",
      "Epoch [27], val_loss1: 0.6202, val_loss2: -0.5933\n",
      "Epoch [28], val_loss1: 0.6181, val_loss2: -0.5908\n",
      "Epoch [29], val_loss1: 0.5986, val_loss2: -0.5685\n",
      "Epoch [30], val_loss1: 0.5946, val_loss2: -0.5663\n",
      "Epoch [31], val_loss1: 0.5918, val_loss2: -0.5649\n",
      "Epoch [32], val_loss1: 0.5810, val_loss2: -0.5553\n",
      "Epoch [33], val_loss1: 0.5354, val_loss2: -0.5104\n",
      "Epoch [34], val_loss1: 0.5385, val_loss2: -0.5133\n",
      "Epoch [35], val_loss1: 0.5121, val_loss2: -0.4878\n",
      "Epoch [36], val_loss1: 0.5291, val_loss2: -0.5063\n",
      "Epoch [37], val_loss1: 0.6000, val_loss2: -0.5752\n",
      "Epoch [38], val_loss1: 0.6003, val_loss2: -0.5775\n",
      "Epoch [39], val_loss1: 0.6001, val_loss2: -0.5786\n"
     ]
    }
   ],
   "source": [
    "history = training(N_EPOCHS,model,train_loader,val_loader)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 295
    },
    "id": "fYwlN0JKVVtN",
    "outputId": "c742ff8b-3b4a-41f5-dd09-effee1be928a"
   },
   "outputs": [
    {
     "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": [
    "plot_history(history)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {
    "id": "ieObNqKYsOzh"
   },
   "outputs": [],
   "source": [
    "torch.save({\n",
    "            'encoder': model.encoder.state_dict(),\n",
    "            'decoder1': model.decoder1.state_dict(),\n",
    "            'decoder2': model.decoder2.state_dict()\n",
    "            }, \"model.pth\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "UsadModel(\n",
       "  (encoder): Encoder(\n",
       "    (linear1): Linear(in_features=1020, out_features=510, bias=True)\n",
       "    (linear2): Linear(in_features=510, out_features=255, bias=True)\n",
       "    (linear3): Linear(in_features=255, out_features=400, bias=True)\n",
       "    (relu): ReLU(inplace=True)\n",
       "  )\n",
       "  (decoder1): Decoder(\n",
       "    (linear1): Linear(in_features=400, out_features=255, bias=True)\n",
       "    (linear2): Linear(in_features=255, out_features=510, bias=True)\n",
       "    (linear3): Linear(in_features=510, out_features=1020, bias=True)\n",
       "    (relu): ReLU(inplace=True)\n",
       "    (sigmoid): Sigmoid()\n",
       "  )\n",
       "  (decoder2): Decoder(\n",
       "    (linear1): Linear(in_features=400, out_features=255, bias=True)\n",
       "    (linear2): Linear(in_features=255, out_features=510, bias=True)\n",
       "    (linear3): Linear(in_features=510, out_features=1020, bias=True)\n",
       "    (relu): ReLU(inplace=True)\n",
       "    (sigmoid): Sigmoid()\n",
       "  )\n",
       ")"
      ]
     },
     "execution_count": 23,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "ymhjbmvR_DgJ"
   },
   "source": [
    "## Testing"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 34
    },
    "id": "b7rbm9wdXKeF",
    "outputId": "076309c7-22be-41f6-f916-5f11cb679672"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<All keys matched successfully>"
      ]
     },
     "execution_count": 24,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "checkpoint = torch.load(\"model.pth\")\n",
    "\n",
    "model.encoder.load_state_dict(checkpoint['encoder'])\n",
    "model.decoder1.load_state_dict(checkpoint['decoder1'])\n",
    "model.decoder2.load_state_dict(checkpoint['decoder2'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {
    "id": "Ry1QTp6V2ny4"
   },
   "outputs": [],
   "source": [
    "results=testing(model,test_loader,0.5,0.5)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "450"
      ]
     },
     "execution_count": 26,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "len(results)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [],
   "source": [
    "windows_labels=[]\n",
    "for i in range(len(labels)-window_size):\n",
    "    windows_labels.append(list(np.int_(labels[i:i+window_size])))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [],
   "source": [
    "y_test = [1.0 if (np.sum(window) > 0) else 0 for window in windows_labels ]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {
    "id": "FSWwxheNvxR7"
   },
   "outputs": [],
   "source": [
    "y_pred=np.concatenate([torch.stack(results[:-1]).flatten().detach().cpu().numpy(),\n",
    "                              results[-1].flatten().detach().cpu().numpy()])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(449899, 449899)"
      ]
     },
     "execution_count": 30,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "len(y_test), len(y_pred)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 279
    },
    "id": "bROUyLM93cG3",
    "outputId": "755359d9-d0fb-4deb-b313-d3c2a2465a26"
   },
   "outputs": [
    {
     "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": [
    "threshold=ROC(y_test,y_pred)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {},
   "outputs": [],
   "source": [
    "y_pred_ = np.zeros(y_pred.shape[0])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {},
   "outputs": [],
   "source": [
    "y_pred_[y_pred >= threshold] = 1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([0., 0., 0., ..., 1., 1., 1.])"
      ]
     },
     "execution_count": 34,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y_pred_"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([0., 0., 0., ..., 0., 0., 0.])"
      ]
     },
     "execution_count": 35,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "np.array(y_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "confusion_matrix(np.array(y_test), y_pred_)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {},
   "outputs": [],
   "source": [
    "import sklearn"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "              precision    recall  f1-score   support\n",
      "\n",
      "         0.0       0.95      0.73      0.82    394613\n",
      "         1.0       0.27      0.73      0.39     55286\n",
      "\n",
      "    accuracy                           0.73    449899\n",
      "   macro avg       0.61      0.73      0.61    449899\n",
      "weighted avg       0.87      0.73      0.77    449899\n",
      "\n"
     ]
    }
   ],
   "source": [
    "print(sklearn.metrics.classification_report(y_test, y_pred_))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "metadata": {},
   "outputs": [],
   "source": [
    "threshold = 10\n",
    "y_pred_ = np.zeros(y_pred.shape[0])\n",
    "y_pred_[y_pred >= threshold] = 1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "              precision    recall  f1-score   support\n",
      "\n",
      "         0.0       0.95      0.99      0.97    394613\n",
      "         1.0       0.90      0.63      0.74     55286\n",
      "\n",
      "    accuracy                           0.95    449899\n",
      "   macro avg       0.92      0.81      0.86    449899\n",
      "weighted avg       0.94      0.95      0.94    449899\n",
      "\n"
     ]
    }
   ],
   "source": [
    "print(sklearn.metrics.classification_report(y_test, y_pred_))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "accelerator": "GPU",
  "colab": {
   "name": "USAD_test.ipynb",
   "provenance": [],
   "toc_visible": true
  },
  "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": "3.8.12"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}