LSTM Autoencoder for Network Anomaly Detection.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "haDl53nO8Oe6"
   },
   "source": [
    "# **Project: LSTM Autoencoder for Network Anomaly Detection**\n",
    "\n",
    "### **Instructions:**\n",
    "\n",
    "In this project, we will create and train an LSTM-based autoencoder to detect anomalies in the KDD99 network traffic dataset. \n",
    "\n",
    "Note that KDD99 does not include timestamps as a feature. The simplest approach to making these discrete datapoints into time-domain data is to assume that each datapoint occurs at the timestep immediately after the previous datapoint. However, more sophisticated approaches can also be adopted (e.g., grouping by TCP connections). The choice of serialization technique (i.e., conversion into time-domain) is up to you.\n",
    "\n",
    "This project must be implemented in a Jupyer Notebook, and must be compatible with Google Colab (i.e., if you are using a particular library that is not on Colab by default, your notebook must install it via !pip install ... ). Your notebook must also contain a section on performance analysis, where you report the performance of your IDS model via performance metrics and/or plots (similar to Section 5).\n",
    "\n",
    "**A very important note:** You should not expect very high detection rates from your model. "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "_2t-Cnfh-ahS"
   },
   "source": [
    "## Read in KDD99 Data Set"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 90
    },
    "id": "3PFGJcB88M6I",
    "outputId": "210a6453-f368-42ea-a527-931ba330869d"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "C:\\Users\\sBumS\\.keras\\datasets\\kddcup.data_10_percent.gz\n"
     ]
    }
   ],
   "source": [
    "#importing required libraries\n",
    "\n",
    "import pandas as pd\n",
    "import numpy as np\n",
    "import tensorflow as tf\n",
    "from tensorflow import keras\n",
    "import pandas as pd\n",
    "import seaborn as sns\n",
    "import matplotlib.pyplot as plt\n",
    "from sklearn import metrics\n",
    "\n",
    "from tensorflow.keras.utils import get_file\n",
    "try:\n",
    "    path = get_file('kddcup.data_10_percent.gz', origin='http://kdd.ics.uci.edu/databases/kddcup99/kddcup.data_10_percent.gz')\n",
    "except:\n",
    "    print('Error downloading')\n",
    "    raise\n",
    "\n",
    "print(path) "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 35
    },
    "id": "HbiUbM0NPOMy",
    "outputId": "e9cd571b-1323-41b2-a2b7-3051b62d3b1b"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Read 494021 rows.\n"
     ]
    }
   ],
   "source": [
    "# This file is a CSV, just no CSV extension or headers\n",
    "# Download from: http://kdd.ics.uci.edu/databases/kddcup99/kddcup99.html\n",
    "df = pd.read_csv(path, header=None)\n",
    "print(\"Read {} rows.\".format(len(df)))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 220
    },
    "id": "MX5AmlE6POBA",
    "outputId": "b969e96f-9a43-4ddf-d6ff-9e148f18a877"
   },
   "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>duration</th>\n",
       "      <th>protocol_type</th>\n",
       "      <th>service</th>\n",
       "      <th>flag</th>\n",
       "      <th>src_bytes</th>\n",
       "      <th>dst_bytes</th>\n",
       "      <th>land</th>\n",
       "      <th>wrong_fragment</th>\n",
       "      <th>urgent</th>\n",
       "      <th>hot</th>\n",
       "      <th>...</th>\n",
       "      <th>dst_host_srv_count</th>\n",
       "      <th>dst_host_same_srv_rate</th>\n",
       "      <th>dst_host_diff_srv_rate</th>\n",
       "      <th>dst_host_same_src_port_rate</th>\n",
       "      <th>dst_host_srv_diff_host_rate</th>\n",
       "      <th>dst_host_serror_rate</th>\n",
       "      <th>dst_host_srv_serror_rate</th>\n",
       "      <th>dst_host_rerror_rate</th>\n",
       "      <th>dst_host_srv_rerror_rate</th>\n",
       "      <th>outcome</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0</td>\n",
       "      <td>tcp</td>\n",
       "      <td>http</td>\n",
       "      <td>SF</td>\n",
       "      <td>181</td>\n",
       "      <td>5450</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>9</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.11</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>normal.</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0</td>\n",
       "      <td>tcp</td>\n",
       "      <td>http</td>\n",
       "      <td>SF</td>\n",
       "      <td>239</td>\n",
       "      <td>486</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>19</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.05</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>normal.</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0</td>\n",
       "      <td>tcp</td>\n",
       "      <td>http</td>\n",
       "      <td>SF</td>\n",
       "      <td>235</td>\n",
       "      <td>1337</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>29</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.03</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>normal.</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0</td>\n",
       "      <td>tcp</td>\n",
       "      <td>http</td>\n",
       "      <td>SF</td>\n",
       "      <td>219</td>\n",
       "      <td>1337</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>39</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.03</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>normal.</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0</td>\n",
       "      <td>tcp</td>\n",
       "      <td>http</td>\n",
       "      <td>SF</td>\n",
       "      <td>217</td>\n",
       "      <td>2032</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>49</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.02</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>normal.</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 42 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "   duration protocol_type service flag  src_bytes  dst_bytes  land  \\\n",
       "0         0           tcp    http   SF        181       5450     0   \n",
       "1         0           tcp    http   SF        239        486     0   \n",
       "2         0           tcp    http   SF        235       1337     0   \n",
       "3         0           tcp    http   SF        219       1337     0   \n",
       "4         0           tcp    http   SF        217       2032     0   \n",
       "\n",
       "   wrong_fragment  urgent  hot  ...  dst_host_srv_count  \\\n",
       "0               0       0    0  ...                   9   \n",
       "1               0       0    0  ...                  19   \n",
       "2               0       0    0  ...                  29   \n",
       "3               0       0    0  ...                  39   \n",
       "4               0       0    0  ...                  49   \n",
       "\n",
       "   dst_host_same_srv_rate  dst_host_diff_srv_rate  \\\n",
       "0                     1.0                     0.0   \n",
       "1                     1.0                     0.0   \n",
       "2                     1.0                     0.0   \n",
       "3                     1.0                     0.0   \n",
       "4                     1.0                     0.0   \n",
       "\n",
       "   dst_host_same_src_port_rate  dst_host_srv_diff_host_rate  \\\n",
       "0                         0.11                          0.0   \n",
       "1                         0.05                          0.0   \n",
       "2                         0.03                          0.0   \n",
       "3                         0.03                          0.0   \n",
       "4                         0.02                          0.0   \n",
       "\n",
       "   dst_host_serror_rate  dst_host_srv_serror_rate  dst_host_rerror_rate  \\\n",
       "0                   0.0                       0.0                   0.0   \n",
       "1                   0.0                       0.0                   0.0   \n",
       "2                   0.0                       0.0                   0.0   \n",
       "3                   0.0                       0.0                   0.0   \n",
       "4                   0.0                       0.0                   0.0   \n",
       "\n",
       "   dst_host_srv_rerror_rate  outcome  \n",
       "0                       0.0  normal.  \n",
       "1                       0.0  normal.  \n",
       "2                       0.0  normal.  \n",
       "3                       0.0  normal.  \n",
       "4                       0.0  normal.  \n",
       "\n",
       "[5 rows x 42 columns]"
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# The CSV file has no column heads, so add them\n",
    "df.columns = ['duration','protocol_type','service','flag','src_bytes','dst_bytes','land','wrong_fragment','urgent','hot',\n",
    "              'num_failed_logins','logged_in','num_compromised','root_shell', 'su_attempted','num_root','num_file_creations','num_shells',\n",
    "              'num_access_files','num_outbound_cmds','is_host_login','is_guest_login','count','srv_count','serror_rate','srv_serror_rate',\n",
    "              'rerror_rate','srv_rerror_rate','same_srv_rate','diff_srv_rate','srv_diff_host_rate','dst_host_count','dst_host_srv_count',\n",
    "              'dst_host_same_srv_rate','dst_host_diff_srv_rate','dst_host_same_src_port_rate','dst_host_srv_diff_host_rate','dst_host_serror_rate',\n",
    "              'dst_host_srv_serror_rate','dst_host_rerror_rate','dst_host_srv_rerror_rate','outcome']\n",
    "\n",
    "# displaying the contents\n",
    "df.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "Tkxa5l5xDLRx"
   },
   "source": [
    "## Data Preprocessing"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 35
    },
    "id": "gNddKIhZPETx",
    "outputId": "7a2e061e-74ea-4514-e87a-65bf2c67de30"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(494021, 42)"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# For now, just drop NA's (rows with missing values)\n",
    "df.dropna(inplace=True,axis=1) \n",
    "df.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 781
    },
    "id": "EY5w2pjiDKA8",
    "outputId": "5499ffbe-82d0-4844-fe0c-862345800ae0"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "duration                         int64\n",
       "protocol_type                   object\n",
       "service                         object\n",
       "flag                            object\n",
       "src_bytes                        int64\n",
       "dst_bytes                        int64\n",
       "land                             int64\n",
       "wrong_fragment                   int64\n",
       "urgent                           int64\n",
       "hot                              int64\n",
       "num_failed_logins                int64\n",
       "logged_in                        int64\n",
       "num_compromised                  int64\n",
       "root_shell                       int64\n",
       "su_attempted                     int64\n",
       "num_root                         int64\n",
       "num_file_creations               int64\n",
       "num_shells                       int64\n",
       "num_access_files                 int64\n",
       "num_outbound_cmds                int64\n",
       "is_host_login                    int64\n",
       "is_guest_login                   int64\n",
       "count                            int64\n",
       "srv_count                        int64\n",
       "serror_rate                    float64\n",
       "srv_serror_rate                float64\n",
       "rerror_rate                    float64\n",
       "srv_rerror_rate                float64\n",
       "same_srv_rate                  float64\n",
       "diff_srv_rate                  float64\n",
       "srv_diff_host_rate             float64\n",
       "dst_host_count                   int64\n",
       "dst_host_srv_count               int64\n",
       "dst_host_same_srv_rate         float64\n",
       "dst_host_diff_srv_rate         float64\n",
       "dst_host_same_src_port_rate    float64\n",
       "dst_host_srv_diff_host_rate    float64\n",
       "dst_host_serror_rate           float64\n",
       "dst_host_srv_serror_rate       float64\n",
       "dst_host_rerror_rate           float64\n",
       "dst_host_srv_rerror_rate       float64\n",
       "outcome                         object\n",
       "dtype: object"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.dtypes"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 461
    },
    "id": "uJH9xUF48Ljr",
    "outputId": "2ee5497e-90c5-48f9-c658-22081ab70eee"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "outcome\n",
       "back.                 2203\n",
       "buffer_overflow.        30\n",
       "ftp_write.               8\n",
       "guess_passwd.           53\n",
       "imap.                   12\n",
       "ipsweep.              1247\n",
       "land.                   21\n",
       "loadmodule.              9\n",
       "multihop.                7\n",
       "neptune.            107201\n",
       "nmap.                  231\n",
       "normal.              97278\n",
       "perl.                    3\n",
       "phf.                     4\n",
       "pod.                   264\n",
       "portsweep.            1040\n",
       "rootkit.                10\n",
       "satan.                1589\n",
       "smurf.              280790\n",
       "spy.                     2\n",
       "teardrop.              979\n",
       "warezclient.          1020\n",
       "warezmaster.            20\n",
       "Name: outcome, dtype: int64"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.groupby('outcome')['outcome'].count()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "49nClouCDBWQ"
   },
   "source": [
    "## Encoding numeric and text data:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "id": "ignXAw75-o1i"
   },
   "outputs": [],
   "source": [
    "# Encode a numeric column as zscores\n",
    "def encode_numeric_zscore(df, name, mean=None, sd=None):\n",
    "    if mean is None:\n",
    "        mean = df[name].mean()\n",
    "\n",
    "    if sd is None:\n",
    "        sd = df[name].std()\n",
    "\n",
    "    df[name] = (df[name] - mean) / sd\n",
    "    \n",
    "# Encode text values to dummy variables(i.e. [1,0,0],[0,1,0],[0,0,1] for red,green,blue)\n",
    "def encode_text_dummy(df, name):\n",
    "    dummies = pd.get_dummies(df[name])\n",
    "    for x in dummies.columns:\n",
    "        dummy_name = f\"{name}-{x}\"\n",
    "        df[dummy_name] = dummies[x]\n",
    "    df.drop(name, axis=1, inplace=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "id": "zRp2aY8AC-uE"
   },
   "outputs": [],
   "source": [
    "#encoding the feature vector\n",
    "text_col =['protocol_type', 'service', 'flag', 'land', 'logged_in', 'is_host_login', 'is_guest_login', ]\n",
    "\n",
    "for i in df.columns:\n",
    "  if i not in text_col:\n",
    "    if i != 'outcome':\n",
    "      encode_numeric_zscore(df, i)\n",
    "\n",
    "for x in text_col:\n",
    "  encode_text_dummy(df, x)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 266
    },
    "id": "-TqbJf4IOKGi",
    "outputId": "fff24138-2278-4e12-b46b-fca92c0fe8f7"
   },
   "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>duration</th>\n",
       "      <th>src_bytes</th>\n",
       "      <th>dst_bytes</th>\n",
       "      <th>wrong_fragment</th>\n",
       "      <th>urgent</th>\n",
       "      <th>hot</th>\n",
       "      <th>num_failed_logins</th>\n",
       "      <th>num_compromised</th>\n",
       "      <th>root_shell</th>\n",
       "      <th>su_attempted</th>\n",
       "      <th>...</th>\n",
       "      <th>flag-S3</th>\n",
       "      <th>flag-SF</th>\n",
       "      <th>flag-SH</th>\n",
       "      <th>land-0</th>\n",
       "      <th>land-1</th>\n",
       "      <th>logged_in-0</th>\n",
       "      <th>logged_in-1</th>\n",
       "      <th>is_host_login-0</th>\n",
       "      <th>is_guest_login-0</th>\n",
       "      <th>is_guest_login-1</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>-0.067792</td>\n",
       "      <td>-0.002879</td>\n",
       "      <td>0.138664</td>\n",
       "      <td>-0.04772</td>\n",
       "      <td>-0.002571</td>\n",
       "      <td>-0.044136</td>\n",
       "      <td>-0.009782</td>\n",
       "      <td>-0.005679</td>\n",
       "      <td>-0.010552</td>\n",
       "      <td>-0.004676</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>-0.067792</td>\n",
       "      <td>-0.002820</td>\n",
       "      <td>-0.011578</td>\n",
       "      <td>-0.04772</td>\n",
       "      <td>-0.002571</td>\n",
       "      <td>-0.044136</td>\n",
       "      <td>-0.009782</td>\n",
       "      <td>-0.005679</td>\n",
       "      <td>-0.010552</td>\n",
       "      <td>-0.004676</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>-0.067792</td>\n",
       "      <td>-0.002824</td>\n",
       "      <td>0.014179</td>\n",
       "      <td>-0.04772</td>\n",
       "      <td>-0.002571</td>\n",
       "      <td>-0.044136</td>\n",
       "      <td>-0.009782</td>\n",
       "      <td>-0.005679</td>\n",
       "      <td>-0.010552</td>\n",
       "      <td>-0.004676</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>-0.067792</td>\n",
       "      <td>-0.002840</td>\n",
       "      <td>0.014179</td>\n",
       "      <td>-0.04772</td>\n",
       "      <td>-0.002571</td>\n",
       "      <td>-0.044136</td>\n",
       "      <td>-0.009782</td>\n",
       "      <td>-0.005679</td>\n",
       "      <td>-0.010552</td>\n",
       "      <td>-0.004676</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>-0.067792</td>\n",
       "      <td>-0.002842</td>\n",
       "      <td>0.035214</td>\n",
       "      <td>-0.04772</td>\n",
       "      <td>-0.002571</td>\n",
       "      <td>-0.044136</td>\n",
       "      <td>-0.009782</td>\n",
       "      <td>-0.005679</td>\n",
       "      <td>-0.010552</td>\n",
       "      <td>-0.004676</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 121 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "   duration  src_bytes  dst_bytes  wrong_fragment    urgent       hot  \\\n",
       "0 -0.067792  -0.002879   0.138664        -0.04772 -0.002571 -0.044136   \n",
       "1 -0.067792  -0.002820  -0.011578        -0.04772 -0.002571 -0.044136   \n",
       "2 -0.067792  -0.002824   0.014179        -0.04772 -0.002571 -0.044136   \n",
       "3 -0.067792  -0.002840   0.014179        -0.04772 -0.002571 -0.044136   \n",
       "4 -0.067792  -0.002842   0.035214        -0.04772 -0.002571 -0.044136   \n",
       "\n",
       "   num_failed_logins  num_compromised  root_shell  su_attempted  ...  flag-S3  \\\n",
       "0          -0.009782        -0.005679   -0.010552     -0.004676  ...        0   \n",
       "1          -0.009782        -0.005679   -0.010552     -0.004676  ...        0   \n",
       "2          -0.009782        -0.005679   -0.010552     -0.004676  ...        0   \n",
       "3          -0.009782        -0.005679   -0.010552     -0.004676  ...        0   \n",
       "4          -0.009782        -0.005679   -0.010552     -0.004676  ...        0   \n",
       "\n",
       "   flag-SF  flag-SH  land-0  land-1  logged_in-0  logged_in-1  \\\n",
       "0        1        0       1       0            0            1   \n",
       "1        1        0       1       0            0            1   \n",
       "2        1        0       1       0            0            1   \n",
       "3        1        0       1       0            0            1   \n",
       "4        1        0       1       0            0            1   \n",
       "\n",
       "   is_host_login-0  is_guest_login-0  is_guest_login-1  \n",
       "0                1                 1                 0  \n",
       "1                1                 1                 0  \n",
       "2                1                 1                 0  \n",
       "3                1                 1                 0  \n",
       "4                1                 1                 0  \n",
       "\n",
       "[5 rows x 121 columns]"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.dropna(inplace=True,axis=1)\n",
    "df[0:5]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 35
    },
    "id": "yTki_IrySDcY",
    "outputId": "a968481a-6032-4322-bbd1-f41a7709c756"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([1, 0], dtype=uint8)"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df['protocol_type-tcp'].unique()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 52
    },
    "id": "WQULPidDTeiP",
    "outputId": "8f5cca59-3a24-48a7-ea37-5ba0c90a9a26"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Normal count: 97278\n",
      "Attack count: 396743\n"
     ]
    }
   ],
   "source": [
    "normal_mask = df['outcome']=='normal.'\n",
    "attack_mask = df['outcome']!='normal.'\n",
    "\n",
    "df.drop('outcome',axis=1,inplace=True)\n",
    "\n",
    "df_normal = df[normal_mask]\n",
    "df_attack = df[attack_mask]\n",
    "\n",
    "print(f\"Normal count: {len(df_normal)}\")\n",
    "print(f\"Attack count: {len(df_attack)}\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {
    "id": "sS_snYlWTgCY"
   },
   "outputs": [],
   "source": [
    "# This is the numeric feature vector, as it goes to the neural net\n",
    "x_normal = df_normal.values\n",
    "x_attack = df_attack.values"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 52
    },
    "id": "HQMsrgUoUfbr",
    "outputId": "24ada8c4-55d3-443b-f309-adea5f76d50a"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Normal train count: 68094\n",
      "Normal test count: 29184\n"
     ]
    }
   ],
   "source": [
    "from sklearn.model_selection import train_test_split\n",
    "\n",
    "x_normal_train, x_normal_test = train_test_split(x_normal, test_size=0.3, random_state=12)\n",
    "\n",
    "print(f\"Normal train count: {len(x_normal_train)}\")\n",
    "print(f\"Normal test count: {len(x_normal_test)}\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 35
    },
    "id": "ijFqQmC9Ul-I",
    "outputId": "d8289d85-9616-4aac-b50b-0030893f2e8f"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "((68094, 120), (29184, 120))"
      ]
     },
     "execution_count": 14,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "x_normal_train.shape, x_normal_test.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {
    "id": "SQLElOnrXdNc"
   },
   "outputs": [],
   "source": [
    "#Scaling the dataset \n",
    "from sklearn.preprocessing import StandardScaler\n",
    "\n",
    "sc = StandardScaler()\n",
    "\n",
    "x_normal_train = sc.fit_transform(x_normal_train)\n",
    "x_normal_test = sc.fit_transform(x_normal_test)\n",
    "x_normal = sc.fit_transform(x_normal)\n",
    "x_attack = sc.fit_transform(x_attack)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "hGz3EXfypegL"
   },
   "source": [
    "We already know that the KDD99 dataset doesn't include timestamps as a feature. So, lets consider simplest approach to making these datapoints into time-domain data i.e., to assume that each datapoint occurs at the timestep immediately after the previous datapoint."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {
    "id": "eiPGMrHOYL57"
   },
   "outputs": [],
   "source": [
    "x_normal_train = pd.DataFrame(x_normal_train)\n",
    "x_normal_test = pd.DataFrame(x_normal_test)\n",
    "x_attack = pd.DataFrame(x_attack)\n",
    "x_normal = pd.DataFrame(x_normal)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {
    "id": "mexyPTQ8Tf8Z"
   },
   "outputs": [],
   "source": [
    "#function to convert to time domain dataset\n",
    "def create_dataset(X, time_steps):\n",
    "    Xs = []\n",
    "    for i in range(len(X) - time_steps):\n",
    "        v = X.iloc[i:(i + time_steps)].values\n",
    "        Xs.append(v)\n",
    "    return np.array(Xs)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 35
    },
    "id": "jGJNP1LKTf2H",
    "outputId": "d960b18c-28cc-47c8-e11e-5c6dff2d66b9"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(68093, 1, 120)\n"
     ]
    }
   ],
   "source": [
    "# reshape to [samples, time_steps, n_features]\n",
    "\n",
    "TIME_STEPS = 1\n",
    "Xnormal_train = create_dataset(x_normal_train, TIME_STEPS)\n",
    "print(Xnormal_train.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 35
    },
    "id": "vgaJUwHYV0HP",
    "outputId": "eb7b9f28-ecfa-4042-ea1d-8a673e960706"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(29183, 1, 120)\n"
     ]
    }
   ],
   "source": [
    "Xnormal_test = create_dataset(x_normal_test, TIME_STEPS)\n",
    "print(Xnormal_test.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 35
    },
    "id": "dNUaZvSczcLE",
    "outputId": "bb9aeee5-2707-4418-dba0-87010b46b1cb"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(396742, 1, 120)\n"
     ]
    }
   ],
   "source": [
    "Xattack = create_dataset(x_attack, TIME_STEPS)\n",
    "print(Xattack.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 35
    },
    "id": "i6Hx1JaVEQpn",
    "outputId": "9ef10339-5116-413b-d1a0-dab68d456b6b"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(396742, 1, 120)\n"
     ]
    }
   ],
   "source": [
    "Xnormal = create_dataset(x_normal,TIME_STEPS)\n",
    "print(Xattack.shape)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "s1LwRkvRjUjU"
   },
   "source": [
    "## LSTM Autoencoder Model Architecture:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 372
    },
    "id": "oDKT-smgCaOu",
    "outputId": "7d2ecaed-593c-4fbf-cd1b-97d06be7a845"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"sequential\"\n",
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "lstm (LSTM)                  (None, 64)                47360     \n",
      "_________________________________________________________________\n",
      "dropout (Dropout)            (None, 64)                0         \n",
      "_________________________________________________________________\n",
      "repeat_vector (RepeatVector) (None, 1, 64)             0         \n",
      "_________________________________________________________________\n",
      "lstm_1 (LSTM)                (None, 1, 32)             12416     \n",
      "_________________________________________________________________\n",
      "dropout_1 (Dropout)          (None, 1, 32)             0         \n",
      "_________________________________________________________________\n",
      "time_distributed (TimeDistri (None, 1, 120)            3960      \n",
      "=================================================================\n",
      "Total params: 63,736\n",
      "Trainable params: 63,736\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "model = keras.Sequential()\n",
    "model.add(keras.layers.LSTM(units=64, input_shape=(Xnormal_train.shape[1], Xnormal_train.shape[2])))\n",
    "model.add(keras.layers.Dropout(rate=0.5))\n",
    "model.add(keras.layers.RepeatVector(n=Xnormal_train.shape[1]))\n",
    "model.add(keras.layers.LSTM(units=32, return_sequences=True))\n",
    "model.add(keras.layers.Dropout(rate=0.2))\n",
    "model.add(keras.layers.TimeDistributed(keras.layers.Dense(units=Xnormal_train.shape[2])))\n",
    "\n",
    "model.compile(loss='mae', optimizer='adam', metrics=['accuracy'])\n",
    "model.summary()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "rwif8JwDh2nY"
   },
   "source": [
    "## LSTM Autoencoder Model Training:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 570
    },
    "id": "R6ppHeDjlEKI",
    "outputId": "ba059b04-b23f-4e3f-e64b-ab76b0003434"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/15\n",
      "426/426 [==============================] - 2s 5ms/step - loss: 0.1052 - accuracy: 0.1657 - val_loss: 0.0610 - val_accuracy: 0.2731\n",
      "Epoch 2/15\n",
      "426/426 [==============================] - 2s 4ms/step - loss: 0.0645 - accuracy: 0.3343 - val_loss: 0.0506 - val_accuracy: 0.3526\n",
      "Epoch 3/15\n",
      "426/426 [==============================] - 2s 4ms/step - loss: 0.0592 - accuracy: 0.3601 - val_loss: 0.0475 - val_accuracy: 0.3027\n",
      "Epoch 4/15\n",
      "426/426 [==============================] - 2s 4ms/step - loss: 0.0559 - accuracy: 0.3752 - val_loss: 0.0461 - val_accuracy: 0.3121\n",
      "Epoch 5/15\n",
      "426/426 [==============================] - 1s 4ms/step - loss: 0.0537 - accuracy: 0.3669 - val_loss: 0.0452 - val_accuracy: 0.3125\n",
      "Epoch 6/15\n",
      "426/426 [==============================] - 2s 4ms/step - loss: 0.0524 - accuracy: 0.3701 - val_loss: 0.0449 - val_accuracy: 0.3141\n",
      "Epoch 7/15\n",
      "426/426 [==============================] - 2s 4ms/step - loss: 0.0515 - accuracy: 0.3745 - val_loss: 0.0444 - val_accuracy: 0.3149\n",
      "Epoch 8/15\n",
      "426/426 [==============================] - 2s 4ms/step - loss: 0.0506 - accuracy: 0.3798 - val_loss: 0.0438 - val_accuracy: 0.3314\n",
      "Epoch 9/15\n",
      "426/426 [==============================] - 2s 4ms/step - loss: 0.0499 - accuracy: 0.3975 - val_loss: 0.0427 - val_accuracy: 0.3337\n",
      "Epoch 10/15\n",
      "426/426 [==============================] - 2s 4ms/step - loss: 0.0490 - accuracy: 0.4408 - val_loss: 0.0414 - val_accuracy: 0.4475\n",
      "Epoch 11/15\n",
      "426/426 [==============================] - 2s 4ms/step - loss: 0.0476 - accuracy: 0.4954 - val_loss: 0.0395 - val_accuracy: 0.4446\n",
      "Epoch 12/15\n",
      "426/426 [==============================] - 2s 4ms/step - loss: 0.0467 - accuracy: 0.5070 - val_loss: 0.0387 - val_accuracy: 0.4447\n",
      "Epoch 13/15\n",
      "426/426 [==============================] - 2s 4ms/step - loss: 0.0462 - accuracy: 0.5095 - val_loss: 0.0383 - val_accuracy: 0.4497\n",
      "Epoch 14/15\n",
      "426/426 [==============================] - 2s 4ms/step - loss: 0.0456 - accuracy: 0.5240 - val_loss: 0.0377 - val_accuracy: 0.4931\n",
      "Epoch 15/15\n",
      "426/426 [==============================] - 2s 4ms/step - loss: 0.0453 - accuracy: 0.5418 - val_loss: 0.0373 - val_accuracy: 0.5036\n"
     ]
    }
   ],
   "source": [
    "history = model.fit(\n",
    "    Xnormal_train, Xnormal_train,\n",
    "    epochs=15,\n",
    "    batch_size=128,\n",
    "    validation_split=0.2, \n",
    "    shuffle = False\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 266
    },
    "id": "aWaAqcwx4Axp",
    "outputId": "10e0292d-5e02-44c0-fabf-e516ad5a5f00"
   },
   "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": [
    "plt.plot(history.history['loss'], label='train')\n",
    "plt.plot(history.history['val_loss'], label='validation')\n",
    "plt.legend();"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "hNCyk6cdiEQy"
   },
   "source": [
    "## Model Evaluation:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 35
    },
    "id": "7nGehivhNKXt",
    "outputId": "bb3d763e-230c-4db2-8bff-dc76dd86c17b"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(68093, 120)"
      ]
     },
     "execution_count": 25,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_train_pred = model.predict(Xnormal_train)\n",
    "\n",
    "train_mae_loss = np.mean(np.abs(X_train_pred - Xnormal_train), axis=1)\n",
    "train_mae_loss.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 268
    },
    "id": "vngf35TONPyE",
    "outputId": "67023318-79a0-4c88-afea-a45a9398af29"
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\sBumS\\anaconda3\\lib\\site-packages\\seaborn\\distributions.py:2551: FutureWarning: `distplot` is a deprecated function and will be removed in a future version. Please adapt your code to use either `displot` (a figure-level function with similar flexibility) or `histplot` (an axes-level function for histograms).\n",
      "  warnings.warn(msg, FutureWarning)\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": [
    "sns.distplot(train_mae_loss, bins=10, kde=True);"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 35
    },
    "id": "H9bfyAXzNUVX",
    "outputId": "83ca021f-d1a7-49eb-c193-f4aa695f2a54"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(29183, 120)"
      ]
     },
     "execution_count": 27,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_test_pred = model.predict(Xnormal_test)\n",
    "\n",
    "test_mae_loss = np.mean(np.abs(X_test_pred - Xnormal_test), axis=1)\n",
    "test_mae_loss.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 265
    },
    "id": "C9Hv-Lp7Ohx0",
    "outputId": "a9e18269-96d9-41ce-f222-0178b68e3991"
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\sBumS\\anaconda3\\lib\\site-packages\\seaborn\\distributions.py:2551: FutureWarning: `distplot` is a deprecated function and will be removed in a future version. Please adapt your code to use either `displot` (a figure-level function with similar flexibility) or `histplot` (an axes-level function for histograms).\n",
      "  warnings.warn(msg, FutureWarning)\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": [
    "sns.distplot(test_mae_loss, bins=10, kde=True);"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 35
    },
    "id": "XGZLWcy-Oq8a",
    "outputId": "15038a70-515b-4ec0-d77d-58ad2f948542"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(97277, 120)"
      ]
     },
     "execution_count": 29,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_normal_pred = model.predict(Xnormal)\n",
    "\n",
    "normal_mae_loss = np.mean(np.abs(X_normal_pred - Xnormal), axis=1)\n",
    "normal_mae_loss.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 265
    },
    "id": "rp1Bqz3SO10M",
    "outputId": "541ec672-4871-4be4-972e-b3428f74406d"
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\sBumS\\anaconda3\\lib\\site-packages\\seaborn\\distributions.py:2551: FutureWarning: `distplot` is a deprecated function and will be removed in a future version. Please adapt your code to use either `displot` (a figure-level function with similar flexibility) or `histplot` (an axes-level function for histograms).\n",
      "  warnings.warn(msg, FutureWarning)\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": [
    "sns.distplot(normal_mae_loss, bins=10, kde=True);"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 35
    },
    "id": "SSByT_hdO_rI",
    "outputId": "47421e55-f33b-4e06-c3e8-7818f93d0b34"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(396742, 120)"
      ]
     },
     "execution_count": 31,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_attack_pred = model.predict(Xattack)\n",
    "\n",
    "attack_mae_loss = np.mean(np.abs(X_attack_pred - Xattack), axis=1)\n",
    "attack_mae_loss.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 265
    },
    "id": "jhxKLUoIPAaO",
    "outputId": "85b6f24b-4e46-4f0b-db9d-366cf3c56c89"
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\sBumS\\anaconda3\\lib\\site-packages\\seaborn\\distributions.py:2551: FutureWarning: `distplot` is a deprecated function and will be removed in a future version. Please adapt your code to use either `displot` (a figure-level function with similar flexibility) or `histplot` (an axes-level function for histograms).\n",
      "  warnings.warn(msg, FutureWarning)\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": [
    "sns.distplot(attack_mae_loss, bins=10, kde=True);"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 230
    },
    "id": "pt7QwhdLJuvu",
    "outputId": "01b5e8b1-bcb5-4bd3-a32d-bd1288ae8793"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "2128/2128 [==============================] - 7s 3ms/step - loss: 0.0369 - accuracy: 0.5075\n",
      "\n",
      " Sample Loss(MAE) & Accuracy Scores (Train): 0.036869991570711136 0.5074677467346191 \n",
      "\n",
      "912/912 [==============================] - 3s 3ms/step - loss: 0.0476 - accuracy: 0.5065\n",
      "\n",
      "Out of Sample Loss(MAE) & Accuracy Scores (Test): 0.047556862235069275 0.5064592361450195 \n",
      "\n",
      "12399/12399 [==============================] - 44s 4ms/step - loss: 0.1694 - accuracy: 0.0017\n",
      "\n",
      "Attack Underway Loss(MAE) & Accuracy Scores (Anomaly): 0.1693747490644455 0.0017391655128449202 \n",
      "\n"
     ]
    }
   ],
   "source": [
    "score1 = model.evaluate(Xnormal_train, Xnormal_train)\n",
    "print(\"\\n Sample Loss(MAE) & Accuracy Scores (Train):\", score1[0], score1[1], \"\\n\") \n",
    "\n",
    "score2 = model.evaluate(Xnormal_test, Xnormal_test)\n",
    "print(\"\\nOut of Sample Loss(MAE) & Accuracy Scores (Test):\", score2[0], score2[1], \"\\n\") \n",
    "\n",
    "score3 = model.evaluate(Xattack, Xattack)\n",
    "print(\"\\nAttack Underway Loss(MAE) & Accuracy Scores (Anomaly):\", score3[0], score3[1], \"\\n\") "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "5SE9ZAAngTZg"
   },
   "source": [
    "**Conclusion:** The accuracy of model prediction on normal train and test sets is almost similarm aroung 0.48. Where as the model accuracy when trained on the anomaly or attack data is 0.0047 which is no way comaprable to the normal data accuracy. \n",
    "\n",
    "It is the same even with the loss, Mean Absolute Error (mae). The normal train and test sets have mae around 0.04 and the anomaly ot attack data has a mae of 0.16. From this we can detect the anomaly if any in the data. "
   ]
  }
 ],
 "metadata": {
  "colab": {
   "collapsed_sections": [],
   "machine_shape": "hm",
   "name": "LSTM Autoencoder for Network Anomaly Detection.ipynb",
   "provenance": [],
   "toc_visible": true
  },
  "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": "3.8.5"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 1
}