notebooks/2-data-preprocessing.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# ML Classification - Network Traffic Analysis\n",
    "## Part 2 - DATA PROCESSING"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "In this notebook, we load the cleaned dataset file and process it into an acceptable format to ML algorithms.\n",
    "\n",
    "> **INPUT:** the cleaned dataset csv file as resulted from the previous cleaning phase.<br>\n",
    "> **OUTPUT:** a processed (final version) of the dataset as a csv file.  "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "***"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 1. INITIALIZATION"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 69,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Import necessary libraries and modules\n",
    "import pandas as pd\n",
    "import numpy as np\n",
    "import seaborn as sns\n",
    "import ipaddress\n",
    "from sklearn.model_selection import train_test_split\n",
    "from sklearn.neighbors import KNeighborsClassifier\n",
    "from sklearn.metrics import accuracy_score, classification_report\n",
    "from sklearn.impute import KNNImputer\n",
    "from sklearn.preprocessing import LabelEncoder, OneHotEncoder\n",
    "from sklearn.preprocessing import MinMaxScaler"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 70,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Set display options\n",
    "pd.set_option('display.max_columns', None)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 2. LOADING CLEANED DATASET"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Reading dataset file into pandas DataFrame"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 71,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Initialize required variables to read the cleaned data file\n",
    "data_file_location = \"..\\\\data\\\\interim\\\\\"\n",
    "data_file_name = \"conn.log.labeled_cleaned\"\n",
    "data_file_ext = \".csv\"\n",
    "\n",
    "# Read the dataset\n",
    "data_df = pd.read_csv(data_file_location + data_file_name + data_file_ext, index_col=0)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Exploring dataset summary and statistics"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 72,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(23145, 15)"
      ]
     },
     "execution_count": 72,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Check dataset shape\n",
    "data_df.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 73,
   "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>id.orig_p</th>\n",
       "      <th>id.resp_p</th>\n",
       "      <th>proto</th>\n",
       "      <th>service</th>\n",
       "      <th>duration</th>\n",
       "      <th>orig_bytes</th>\n",
       "      <th>resp_bytes</th>\n",
       "      <th>conn_state</th>\n",
       "      <th>missed_bytes</th>\n",
       "      <th>history</th>\n",
       "      <th>orig_pkts</th>\n",
       "      <th>orig_ip_bytes</th>\n",
       "      <th>resp_pkts</th>\n",
       "      <th>resp_ip_bytes</th>\n",
       "      <th>label</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>41040</td>\n",
       "      <td>80</td>\n",
       "      <td>tcp</td>\n",
       "      <td>NaN</td>\n",
       "      <td>3.139211</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>S0</td>\n",
       "      <td>0</td>\n",
       "      <td>S</td>\n",
       "      <td>3</td>\n",
       "      <td>180</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>Benign</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>41040</td>\n",
       "      <td>80</td>\n",
       "      <td>tcp</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>S0</td>\n",
       "      <td>0</td>\n",
       "      <td>S</td>\n",
       "      <td>1</td>\n",
       "      <td>60</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>Benign</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>41040</td>\n",
       "      <td>80</td>\n",
       "      <td>tcp</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>S0</td>\n",
       "      <td>0</td>\n",
       "      <td>S</td>\n",
       "      <td>1</td>\n",
       "      <td>60</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>Benign</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>41040</td>\n",
       "      <td>80</td>\n",
       "      <td>tcp</td>\n",
       "      <td>http</td>\n",
       "      <td>1.477656</td>\n",
       "      <td>149.0</td>\n",
       "      <td>128252.0</td>\n",
       "      <td>SF</td>\n",
       "      <td>2896</td>\n",
       "      <td>ShADadttcfF</td>\n",
       "      <td>94</td>\n",
       "      <td>5525</td>\n",
       "      <td>96</td>\n",
       "      <td>139044</td>\n",
       "      <td>Benign</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>41042</td>\n",
       "      <td>80</td>\n",
       "      <td>tcp</td>\n",
       "      <td>NaN</td>\n",
       "      <td>3.147116</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>S0</td>\n",
       "      <td>0</td>\n",
       "      <td>S</td>\n",
       "      <td>3</td>\n",
       "      <td>180</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>Benign</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   id.orig_p  id.resp_p proto service  duration  orig_bytes  resp_bytes  \\\n",
       "0      41040         80   tcp     NaN  3.139211         0.0         0.0   \n",
       "1      41040         80   tcp     NaN       NaN         NaN         NaN   \n",
       "2      41040         80   tcp     NaN       NaN         NaN         NaN   \n",
       "3      41040         80   tcp    http  1.477656       149.0    128252.0   \n",
       "4      41042         80   tcp     NaN  3.147116         0.0         0.0   \n",
       "\n",
       "  conn_state  missed_bytes      history  orig_pkts  orig_ip_bytes  resp_pkts  \\\n",
       "0         S0             0            S          3            180          0   \n",
       "1         S0             0            S          1             60          0   \n",
       "2         S0             0            S          1             60          0   \n",
       "3         SF          2896  ShADadttcfF         94           5525         96   \n",
       "4         S0             0            S          3            180          0   \n",
       "\n",
       "   resp_ip_bytes   label  \n",
       "0              0  Benign  \n",
       "1              0  Benign  \n",
       "2              0  Benign  \n",
       "3         139044  Benign  \n",
       "4              0  Benign  "
      ]
     },
     "execution_count": 73,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Check dataset head\n",
    "data_df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 74,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<class 'pandas.core.frame.DataFrame'>\n",
      "Int64Index: 23145 entries, 0 to 23144\n",
      "Data columns (total 15 columns):\n",
      " #   Column         Non-Null Count  Dtype  \n",
      "---  ------         --------------  -----  \n",
      " 0   id.orig_p      23145 non-null  int64  \n",
      " 1   id.resp_p      23145 non-null  int64  \n",
      " 2   proto          23145 non-null  object \n",
      " 3   service        1847 non-null   object \n",
      " 4   duration       5321 non-null   float64\n",
      " 5   orig_bytes     5321 non-null   float64\n",
      " 6   resp_bytes     5321 non-null   float64\n",
      " 7   conn_state     23145 non-null  object \n",
      " 8   missed_bytes   23145 non-null  int64  \n",
      " 9   history        23145 non-null  object \n",
      " 10  orig_pkts      23145 non-null  int64  \n",
      " 11  orig_ip_bytes  23145 non-null  int64  \n",
      " 12  resp_pkts      23145 non-null  int64  \n",
      " 13  resp_ip_bytes  23145 non-null  int64  \n",
      " 14  label          23145 non-null  object \n",
      "dtypes: float64(3), int64(7), object(5)\n",
      "memory usage: 2.8+ MB\n"
     ]
    }
   ],
   "source": [
    "# Check dataset summary\n",
    "data_df.info()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 3. DATA PREPROCESSING"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Analyzing target attribute"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 75,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0"
      ]
     },
     "execution_count": 75,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Check null values in the target attribute\n",
    "data_df[\"label\"].isna().sum()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 76,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Malicious    21222\n",
       "Benign        1923\n",
       "Name: label, dtype: int64"
      ]
     },
     "execution_count": 76,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Check values distribution\n",
    "data_df[\"label\"].value_counts()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 77,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<Axes: xlabel='label', ylabel='count'>"
      ]
     },
     "execution_count": 77,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Plot target attribute on a count plot\n",
    "sns.countplot(data=data_df, x=\"label\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "- As we can see from these statistics, the target attribute is highly imbalanced.\n",
    "- We will maintain the original distribution for now until we explore the models' performance. \n",
    "- Suitable techniques will be applied to re-balance the labels if we couldn't achieve an acceptable prediction accuracy. "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Encoding target attribute"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 78,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{'Benign': 0, 'Malicious': 1}"
      ]
     },
     "execution_count": 78,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Initialize encoder with default parameters\n",
    "target_le = LabelEncoder()\n",
    "\n",
    "# Fit the encoder to the target attribute\n",
    "encoded_attribute = target_le.fit_transform(data_df[\"label\"])\n",
    "\n",
    "# Replace target attribute with encoded values\n",
    "data_df[\"label\"] = encoded_attribute\n",
    "\n",
    "# Check mapped labels\n",
    "dict(zip(target_le.classes_, target_le.transform(target_le.classes_)))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Handling outliers"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 79,
   "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>duration</th>\n",
       "      <th>orig_bytes</th>\n",
       "      <th>resp_bytes</th>\n",
       "      <th>missed_bytes</th>\n",
       "      <th>orig_pkts</th>\n",
       "      <th>orig_ip_bytes</th>\n",
       "      <th>resp_pkts</th>\n",
       "      <th>resp_ip_bytes</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>count</th>\n",
       "      <td>5321.000000</td>\n",
       "      <td>5.321000e+03</td>\n",
       "      <td>5321.000000</td>\n",
       "      <td>23145.000000</td>\n",
       "      <td>23145.000000</td>\n",
       "      <td>2.314500e+04</td>\n",
       "      <td>23145.000000</td>\n",
       "      <td>23145.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>mean</th>\n",
       "      <td>22.806503</td>\n",
       "      <td>1.478868e+04</td>\n",
       "      <td>350.429431</td>\n",
       "      <td>2.127112</td>\n",
       "      <td>6.375157</td>\n",
       "      <td>3.664312e+03</td>\n",
       "      <td>0.611017</td>\n",
       "      <td>111.218967</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>std</th>\n",
       "      <td>722.522302</td>\n",
       "      <td>1.036441e+06</td>\n",
       "      <td>5378.262771</td>\n",
       "      <td>102.490787</td>\n",
       "      <td>178.548725</td>\n",
       "      <td>5.003762e+05</td>\n",
       "      <td>8.305898</td>\n",
       "      <td>2713.082822</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>min</th>\n",
       "      <td>0.000497</td>\n",
       "      <td>0.000000e+00</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000e+00</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25%</th>\n",
       "      <td>2.075814</td>\n",
       "      <td>0.000000e+00</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000e+00</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>50%</th>\n",
       "      <td>3.110974</td>\n",
       "      <td>0.000000e+00</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000e+00</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>75%</th>\n",
       "      <td>3.153695</td>\n",
       "      <td>6.200000e+01</td>\n",
       "      <td>243.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>7.600000e+01</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>max</th>\n",
       "      <td>48976.819063</td>\n",
       "      <td>7.554662e+07</td>\n",
       "      <td>164266.000000</td>\n",
       "      <td>5792.000000</td>\n",
       "      <td>18444.000000</td>\n",
       "      <td>7.606306e+07</td>\n",
       "      <td>1070.000000</td>\n",
       "      <td>168910.000000</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "           duration    orig_bytes     resp_bytes  missed_bytes     orig_pkts  \\\n",
       "count   5321.000000  5.321000e+03    5321.000000  23145.000000  23145.000000   \n",
       "mean      22.806503  1.478868e+04     350.429431      2.127112      6.375157   \n",
       "std      722.522302  1.036441e+06    5378.262771    102.490787    178.548725   \n",
       "min        0.000497  0.000000e+00       0.000000      0.000000      0.000000   \n",
       "25%        2.075814  0.000000e+00       0.000000      0.000000      0.000000   \n",
       "50%        3.110974  0.000000e+00       0.000000      0.000000      0.000000   \n",
       "75%        3.153695  6.200000e+01     243.000000      0.000000      1.000000   \n",
       "max    48976.819063  7.554662e+07  164266.000000   5792.000000  18444.000000   \n",
       "\n",
       "       orig_ip_bytes     resp_pkts  resp_ip_bytes  \n",
       "count   2.314500e+04  23145.000000   23145.000000  \n",
       "mean    3.664312e+03      0.611017     111.218967  \n",
       "std     5.003762e+05      8.305898    2713.082822  \n",
       "min     0.000000e+00      0.000000       0.000000  \n",
       "25%     0.000000e+00      0.000000       0.000000  \n",
       "50%     0.000000e+00      0.000000       0.000000  \n",
       "75%     7.600000e+01      0.000000       0.000000  \n",
       "max     7.606306e+07   1070.000000  168910.000000  "
      ]
     },
     "execution_count": 79,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Use describe() method to obtain general statistics about the numerical features\n",
    "numerical_features = [\"duration\", \"orig_bytes\",\t\"resp_bytes\", \"missed_bytes\", \"orig_pkts\",\t\"orig_ip_bytes\", \"resp_pkts\", \"resp_ip_bytes\"]\n",
    "data_df[numerical_features].describe()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "- In a quick reading of the statistical summary, we noticed many features with values that can be seen as outliers. \n",
    "- It's fairly difficult to confirm such values and their actual reasons without proper explanations by experts in the corresponding domain.\n",
    "- For the sake of this demo, we will treat only \"duration\" as a feature with actual outliers as it's the only feature with a significant amount of unique values."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 80,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<Axes: ylabel='duration'>"
      ]
     },
     "execution_count": 80,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Plot \"duration\" feature on a boxplot\n",
    "sns.boxplot(data=data_df, y=\"duration\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 81,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(min,max) bounds for \"duration\": (0.4589924999999997,4.7705165)\n"
     ]
    }
   ],
   "source": [
    "# Replace outliers using IQR (Inter-quartile Range)\n",
    "outliers_columns = ['duration']\n",
    "for col_name in outliers_columns:\n",
    "    # Calculate first and third quartiles\n",
    "    q1, q3 = np.nanpercentile(data_df[col_name],[25,75])\n",
    "    \n",
    "    # Calculate the inter-quartile range\n",
    "    intr_qr = q3-q1\n",
    "\n",
    "    # Calculate lower and higher bounds\n",
    "    iqr_min_val = q1-(1.5*intr_qr)\n",
    "    iqr_max_val = q3+(1.5*intr_qr)\n",
    "    print(f\"(min,max) bounds for \\\"{col_name}\\\": ({iqr_min_val},{iqr_max_val})\")\n",
    "    \n",
    "    # Replace values that are less than min or larger then max with np.nan\n",
    "    data_df.loc[data_df[col_name] < iqr_min_val, col_name] = np.nan\n",
    "    data_df.loc[data_df[col_name] > iqr_max_val, col_name] = np.nan"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 82,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "count    3718.000000\n",
       "mean        2.848130\n",
       "std         0.806614\n",
       "min         0.553685\n",
       "25%         3.085004\n",
       "50%         3.110717\n",
       "75%         3.140398\n",
       "max         4.723553\n",
       "Name: duration, dtype: float64"
      ]
     },
     "execution_count": 82,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Reevaluate the new distribution of values\n",
    "data_df[\"duration\"].describe()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Handling missing values"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 83,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "service          21298\n",
       "duration         19427\n",
       "orig_bytes       17824\n",
       "resp_bytes       17824\n",
       "id.orig_p            0\n",
       "id.resp_p            0\n",
       "proto                0\n",
       "conn_state           0\n",
       "missed_bytes         0\n",
       "history              0\n",
       "orig_pkts            0\n",
       "orig_ip_bytes        0\n",
       "resp_pkts            0\n",
       "resp_ip_bytes        0\n",
       "label                0\n",
       "dtype: int64"
      ]
     },
     "execution_count": 83,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Check the number of null values in each column\n",
    "data_df.isnull().sum().sort_values(ascending=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 84,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<Axes: >"
      ]
     },
     "execution_count": 84,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Check null values using heatmap\n",
    "sns.heatmap(data=data_df.isnull(), yticklabels=False, cbar=False, cmap=\"viridis\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "- There are only 4 columns with a lot of missing values, one categorical (\"service) and three numerical (\"duration\", \"orig_bytes\", \"resp_bytes\").\n",
    "- One possible approach is to entirely discard these columns from the dataset. \n",
    "- Instead, we will try to impute their values to preserve any hidden information and check if we can obtain better prediction performance."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "##### Impute missing values: categorical features"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 85,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<Axes: xlabel='label', ylabel='count'>"
      ]
     },
     "execution_count": 85,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Check the relationship between the \"service\" and the target attribute\n",
    "sns.countplot(data=data_df, x=\"label\", hue=\"service\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "- Due to this visualization, almost all *Malicious* observations belong to a specific service type (\"irc\"), while the other 3 types represent *Benign* samples.\n",
    "- Accordingly, although \"service\" feature has high number of missing values, it has a strong correlation with the target attribute.\n",
    "- The decision here is to keep \"service\" attribute while trying to impute its missing values using another classifier. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 86,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Prediction accuracy for 'service' is: 1.0\n",
      "Classification report:\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "         dns       1.00      1.00      1.00        41\n",
      "        http       1.00      1.00      1.00         5\n",
      "         irc       1.00      1.00      1.00       324\n",
      "\n",
      "    accuracy                           1.00       370\n",
      "   macro avg       1.00      1.00      1.00       370\n",
      "weighted avg       1.00      1.00      1.00       370\n",
      "\n"
     ]
    }
   ],
   "source": [
    "# Select specific columns to be used for the classification, here we initially select the numerical attributes with no missing values.\n",
    "srv_training_columns = [\"id.orig_p\",\"id.resp_p\",\"missed_bytes\",\"orig_pkts\",\"orig_ip_bytes\",\"resp_pkts\",\"resp_ip_bytes\"] \n",
    "\n",
    "# Split the rows into two datasets containing rows with/without \"service\"\n",
    "data_df_with_service = data_df[data_df[\"service\"].notna()]\n",
    "data_df_no_service = data_df[data_df[\"service\"].isna()]\n",
    "\n",
    "# Split the service dataset into dependent and independent features\n",
    "srv_X = data_df_with_service[srv_training_columns]\n",
    "srv_y = data_df_with_service[\"service\"].values\n",
    "\n",
    "# Split into train/test subsets\n",
    "srv_X_train, srv_X_test, srv_y_train, srv_y_test = train_test_split(srv_X, srv_y, test_size=0.2, random_state=0)\n",
    "\n",
    "# Create KNN estimator and fit it\n",
    "srv_knn = KNeighborsClassifier(n_neighbors=3)\n",
    "srv_knn.fit(srv_X_train, srv_y_train)\n",
    "\n",
    "# Predict missing values\n",
    "srv_y_pred = srv_knn.predict(srv_X_test)\n",
    "\n",
    "# Check predictions accuracy\n",
    "srv_accuracy_test = accuracy_score(srv_y_test, srv_y_pred)\n",
    "print(f\"Prediction accuracy for 'service' is: {srv_accuracy_test}\")\n",
    "print(\"Classification report:\")\n",
    "print(classification_report(srv_y_test, srv_y_pred))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The classification model achieved an accuracy of 100%, which means all samples in the test subset were correctly predicted. Now we can use this model to predict missing \"service\" fields."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 87,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Predict \"service\" for missing values\n",
    "srv_predictions = srv_knn.predict(data_df_no_service[srv_training_columns])\n",
    "\n",
    "# Update the original data set with predicted \"service\" values\n",
    "data_df.loc[data_df[\"service\"].isna(), \"service\"] = srv_predictions"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "##### Impute missing values: numerical features"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 88,
   "metadata": {},
   "outputs": [],
   "source": [
    "# To preserve hidden correlations with other features in the dataset, we will use a KNN imputer to estimate the missing values based on relationships with other numerical features.\n",
    "numerical_features = data_df.drop(\"label\", axis=1).select_dtypes(include=\"number\").columns\n",
    "knn_imputer = KNNImputer()\n",
    "data_df_after_imputing = knn_imputer.fit_transform(data_df[numerical_features])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 89,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Update original data set to fill missing values with imputed ones\n",
    "data_df[numerical_features] = data_df_after_imputing"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 90,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "id.orig_p        0\n",
       "id.resp_p        0\n",
       "proto            0\n",
       "service          0\n",
       "duration         0\n",
       "orig_bytes       0\n",
       "resp_bytes       0\n",
       "conn_state       0\n",
       "missed_bytes     0\n",
       "history          0\n",
       "orig_pkts        0\n",
       "orig_ip_bytes    0\n",
       "resp_pkts        0\n",
       "resp_ip_bytes    0\n",
       "label            0\n",
       "dtype: int64"
      ]
     },
     "execution_count": 90,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Confirm all missing values were successfully imputed\n",
    "data_df.isnull().sum().sort_values(ascending=False)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Scaling numerical attributes"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "As we aim to compare several classifiers, and since some of them rely on distance-based comparisons, we will scale the numerical features to have them represented in a unified distribution."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 91,
   "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>id.orig_p</th>\n",
       "      <th>id.resp_p</th>\n",
       "      <th>duration</th>\n",
       "      <th>orig_bytes</th>\n",
       "      <th>resp_bytes</th>\n",
       "      <th>missed_bytes</th>\n",
       "      <th>orig_pkts</th>\n",
       "      <th>orig_ip_bytes</th>\n",
       "      <th>resp_pkts</th>\n",
       "      <th>resp_ip_bytes</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>count</th>\n",
       "      <td>23145.000000</td>\n",
       "      <td>23145.000000</td>\n",
       "      <td>23145.000000</td>\n",
       "      <td>2.314500e+04</td>\n",
       "      <td>23145.000000</td>\n",
       "      <td>23145.000000</td>\n",
       "      <td>23145.000000</td>\n",
       "      <td>2.314500e+04</td>\n",
       "      <td>23145.000000</td>\n",
       "      <td>23145.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>mean</th>\n",
       "      <td>56409.651847</td>\n",
       "      <td>2324.573558</td>\n",
       "      <td>3.019661</td>\n",
       "      <td>3.435934e+03</td>\n",
       "      <td>87.489721</td>\n",
       "      <td>2.127112</td>\n",
       "      <td>6.375157</td>\n",
       "      <td>3.664312e+03</td>\n",
       "      <td>0.611017</td>\n",
       "      <td>111.218967</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>std</th>\n",
       "      <td>17176.590987</td>\n",
       "      <td>5362.933817</td>\n",
       "      <td>0.362171</td>\n",
       "      <td>4.969529e+05</td>\n",
       "      <td>2582.581881</td>\n",
       "      <td>102.490787</td>\n",
       "      <td>178.548725</td>\n",
       "      <td>5.003762e+05</td>\n",
       "      <td>8.305898</td>\n",
       "      <td>2713.082822</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>min</th>\n",
       "      <td>0.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.553685</td>\n",
       "      <td>0.000000e+00</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000e+00</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25%</th>\n",
       "      <td>53622.000000</td>\n",
       "      <td>80.000000</td>\n",
       "      <td>3.093835</td>\n",
       "      <td>4.800000e+01</td>\n",
       "      <td>7.800000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000e+00</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>50%</th>\n",
       "      <td>65279.000000</td>\n",
       "      <td>80.000000</td>\n",
       "      <td>3.093835</td>\n",
       "      <td>5.460000e+01</td>\n",
       "      <td>7.800000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000e+00</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>75%</th>\n",
       "      <td>65279.000000</td>\n",
       "      <td>6667.000000</td>\n",
       "      <td>3.093835</td>\n",
       "      <td>5.460000e+01</td>\n",
       "      <td>7.800000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>7.600000e+01</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>max</th>\n",
       "      <td>65279.000000</td>\n",
       "      <td>63798.000000</td>\n",
       "      <td>4.723553</td>\n",
       "      <td>7.554662e+07</td>\n",
       "      <td>164266.000000</td>\n",
       "      <td>5792.000000</td>\n",
       "      <td>18444.000000</td>\n",
       "      <td>7.606306e+07</td>\n",
       "      <td>1070.000000</td>\n",
       "      <td>168910.000000</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "          id.orig_p     id.resp_p      duration    orig_bytes     resp_bytes  \\\n",
       "count  23145.000000  23145.000000  23145.000000  2.314500e+04   23145.000000   \n",
       "mean   56409.651847   2324.573558      3.019661  3.435934e+03      87.489721   \n",
       "std    17176.590987   5362.933817      0.362171  4.969529e+05    2582.581881   \n",
       "min        0.000000      1.000000      0.553685  0.000000e+00       0.000000   \n",
       "25%    53622.000000     80.000000      3.093835  4.800000e+01       7.800000   \n",
       "50%    65279.000000     80.000000      3.093835  5.460000e+01       7.800000   \n",
       "75%    65279.000000   6667.000000      3.093835  5.460000e+01       7.800000   \n",
       "max    65279.000000  63798.000000      4.723553  7.554662e+07  164266.000000   \n",
       "\n",
       "       missed_bytes     orig_pkts  orig_ip_bytes     resp_pkts  resp_ip_bytes  \n",
       "count  23145.000000  23145.000000   2.314500e+04  23145.000000   23145.000000  \n",
       "mean       2.127112      6.375157   3.664312e+03      0.611017     111.218967  \n",
       "std      102.490787    178.548725   5.003762e+05      8.305898    2713.082822  \n",
       "min        0.000000      0.000000   0.000000e+00      0.000000       0.000000  \n",
       "25%        0.000000      0.000000   0.000000e+00      0.000000       0.000000  \n",
       "50%        0.000000      0.000000   0.000000e+00      0.000000       0.000000  \n",
       "75%        0.000000      1.000000   7.600000e+01      0.000000       0.000000  \n",
       "max     5792.000000  18444.000000   7.606306e+07   1070.000000  168910.000000  "
      ]
     },
     "execution_count": 91,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Check statistics for numerical features\n",
    "numerical_features = [\"id.orig_p\", \"id.resp_p\", \"duration\", \"orig_bytes\", \"resp_bytes\", \"missed_bytes\", \"orig_pkts\", \"orig_ip_bytes\",\t\"resp_pkts\", \"resp_ip_bytes\"]\n",
    "data_df[numerical_features].describe()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Since most attributes have no normal distribution, it's more suitable to apply a normalization (between 0 and 1, using MinMaxScaler) instead of standardization."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 92,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Initialize and apply MinMaxScaler scaler\n",
    "min_max_scaler = MinMaxScaler()\n",
    "data_df[numerical_features] = min_max_scaler.fit_transform(data_df[numerical_features])\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 93,
   "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>id.orig_p</th>\n",
       "      <th>id.resp_p</th>\n",
       "      <th>duration</th>\n",
       "      <th>orig_bytes</th>\n",
       "      <th>resp_bytes</th>\n",
       "      <th>missed_bytes</th>\n",
       "      <th>orig_pkts</th>\n",
       "      <th>orig_ip_bytes</th>\n",
       "      <th>resp_pkts</th>\n",
       "      <th>resp_ip_bytes</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>count</th>\n",
       "      <td>23145.000000</td>\n",
       "      <td>23145.000000</td>\n",
       "      <td>23145.000000</td>\n",
       "      <td>2.314500e+04</td>\n",
       "      <td>23145.000000</td>\n",
       "      <td>23145.000000</td>\n",
       "      <td>23145.000000</td>\n",
       "      <td>2.314500e+04</td>\n",
       "      <td>23145.000000</td>\n",
       "      <td>23145.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>mean</th>\n",
       "      <td>0.864132</td>\n",
       "      <td>0.036421</td>\n",
       "      <td>0.591380</td>\n",
       "      <td>4.548098e-05</td>\n",
       "      <td>0.000533</td>\n",
       "      <td>0.000367</td>\n",
       "      <td>0.000346</td>\n",
       "      <td>4.817466e-05</td>\n",
       "      <td>0.000571</td>\n",
       "      <td>0.000658</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>std</th>\n",
       "      <td>0.263126</td>\n",
       "      <td>0.084062</td>\n",
       "      <td>0.086854</td>\n",
       "      <td>6.578095e-03</td>\n",
       "      <td>0.015722</td>\n",
       "      <td>0.017695</td>\n",
       "      <td>0.009681</td>\n",
       "      <td>6.578439e-03</td>\n",
       "      <td>0.007763</td>\n",
       "      <td>0.016062</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>min</th>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000e+00</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000e+00</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25%</th>\n",
       "      <td>0.821428</td>\n",
       "      <td>0.001238</td>\n",
       "      <td>0.609168</td>\n",
       "      <td>6.353692e-07</td>\n",
       "      <td>0.000047</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000e+00</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>50%</th>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.001238</td>\n",
       "      <td>0.609168</td>\n",
       "      <td>7.227325e-07</td>\n",
       "      <td>0.000047</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000e+00</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>75%</th>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.104488</td>\n",
       "      <td>0.609168</td>\n",
       "      <td>7.227325e-07</td>\n",
       "      <td>0.000047</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000054</td>\n",
       "      <td>9.991710e-07</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>max</th>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000e+00</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000e+00</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>1.000000</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "          id.orig_p     id.resp_p      duration    orig_bytes    resp_bytes  \\\n",
       "count  23145.000000  23145.000000  23145.000000  2.314500e+04  23145.000000   \n",
       "mean       0.864132      0.036421      0.591380  4.548098e-05      0.000533   \n",
       "std        0.263126      0.084062      0.086854  6.578095e-03      0.015722   \n",
       "min        0.000000      0.000000      0.000000  0.000000e+00      0.000000   \n",
       "25%        0.821428      0.001238      0.609168  6.353692e-07      0.000047   \n",
       "50%        1.000000      0.001238      0.609168  7.227325e-07      0.000047   \n",
       "75%        1.000000      0.104488      0.609168  7.227325e-07      0.000047   \n",
       "max        1.000000      1.000000      1.000000  1.000000e+00      1.000000   \n",
       "\n",
       "       missed_bytes     orig_pkts  orig_ip_bytes     resp_pkts  resp_ip_bytes  \n",
       "count  23145.000000  23145.000000   2.314500e+04  23145.000000   23145.000000  \n",
       "mean       0.000367      0.000346   4.817466e-05      0.000571       0.000658  \n",
       "std        0.017695      0.009681   6.578439e-03      0.007763       0.016062  \n",
       "min        0.000000      0.000000   0.000000e+00      0.000000       0.000000  \n",
       "25%        0.000000      0.000000   0.000000e+00      0.000000       0.000000  \n",
       "50%        0.000000      0.000000   0.000000e+00      0.000000       0.000000  \n",
       "75%        0.000000      0.000054   9.991710e-07      0.000000       0.000000  \n",
       "max        1.000000      1.000000   1.000000e+00      1.000000       1.000000  "
      ]
     },
     "execution_count": 93,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Check statistics for scaled features\n",
    "data_df[numerical_features].describe()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Encoding categorical features"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 94,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<class 'pandas.core.frame.DataFrame'>\n",
      "Int64Index: 23145 entries, 0 to 23144\n",
      "Data columns (total 15 columns):\n",
      " #   Column         Non-Null Count  Dtype  \n",
      "---  ------         --------------  -----  \n",
      " 0   id.orig_p      23145 non-null  float64\n",
      " 1   id.resp_p      23145 non-null  float64\n",
      " 2   proto          23145 non-null  object \n",
      " 3   service        23145 non-null  object \n",
      " 4   duration       23145 non-null  float64\n",
      " 5   orig_bytes     23145 non-null  float64\n",
      " 6   resp_bytes     23145 non-null  float64\n",
      " 7   conn_state     23145 non-null  object \n",
      " 8   missed_bytes   23145 non-null  float64\n",
      " 9   history        23145 non-null  object \n",
      " 10  orig_pkts      23145 non-null  float64\n",
      " 11  orig_ip_bytes  23145 non-null  float64\n",
      " 12  resp_pkts      23145 non-null  float64\n",
      " 13  resp_ip_bytes  23145 non-null  float64\n",
      " 14  label          23145 non-null  int32  \n",
      "dtypes: float64(10), int32(1), object(4)\n",
      "memory usage: 2.7+ MB\n"
     ]
    }
   ],
   "source": [
    "# Look for categorical features in the dataset\n",
    "data_df.info()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "- Since all categorical features don't imply an ordered relationship between their values, they can be encoded using One-Hot Encoding.\n",
    "- We need first to check features with rare values and map them to \"other\", in order to avoid sparse columns with statistically negligible impact and higher computational effect."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "##### Encoding categorical features: handle rare values"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 95,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Column (proto) has (2) distinct values.\n",
      "Column (service) has (4) distinct values.\n",
      "Column (conn_state) has (6) distinct values.\n",
      "Column (history) has (26) distinct values.\n"
     ]
    }
   ],
   "source": [
    "# Check the number of unique values in each feature\n",
    "categorical_features = [\"proto\",\"service\",\"conn_state\",\"history\"]\n",
    "for c in categorical_features:\n",
    "    print(f\"Column ({c}) has ({data_df[c].nunique()}) distinct values.\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 96,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "C               14252\n",
       "S                5417\n",
       "ShAdDaf          1477\n",
       "D                 978\n",
       "Dd                836\n",
       "ShAdDaft          102\n",
       "ShAdfDr            48\n",
       "CCCC                6\n",
       "ShADadtcfF          3\n",
       "ShADadttcfF         3\n",
       "ShAdDatfr           2\n",
       "CCC                 2\n",
       "ShADadf             2\n",
       "ShDadAf             2\n",
       "ShAfdtDr            2\n",
       "ShADacdtfF          2\n",
       "ShADadtctfF         2\n",
       "ShAdDatf            1\n",
       "ShADadttfF          1\n",
       "ShAdD               1\n",
       "ShADadtctfFR        1\n",
       "ShAdDfr             1\n",
       "ShAD                1\n",
       "DdAtaFf             1\n",
       "ShADad              1\n",
       "ShAdDa              1\n",
       "Name: history, dtype: int64"
      ]
     },
     "execution_count": 96,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Check values of \"history\" because it has too many unique values.\n",
    "history_val_counts = data_df[\"history\"].value_counts()\n",
    "history_val_counts"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 97,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "C           14252\n",
       "S            5417\n",
       "ShAdDaf      1477\n",
       "D             978\n",
       "Dd            836\n",
       "ShAdDaft      102\n",
       "ShAdfDr        48\n",
       "Other          35\n",
       "Name: history, dtype: int64"
      ]
     },
     "execution_count": 97,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Map values to their frequencies\n",
    "history_freq_map = data_df[\"history\"].map(history_val_counts)\n",
    "\n",
    "# Replace low frequent values in \"history\" with \"Other\" using the corresponding frequency map\n",
    "data_df[\"history\"] = data_df[\"history\"].mask(history_freq_map < 10, \"Other\")\n",
    "\n",
    "# Check \"history\" values after mapping\n",
    "data_df[\"history\"].value_counts()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "##### Encoding categorical features: apply one-hot encoder"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 98,
   "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>proto_tcp</th>\n",
       "      <th>proto_udp</th>\n",
       "      <th>service_dhcp</th>\n",
       "      <th>service_dns</th>\n",
       "      <th>service_http</th>\n",
       "      <th>service_irc</th>\n",
       "      <th>conn_state_OTH</th>\n",
       "      <th>conn_state_RSTR</th>\n",
       "      <th>conn_state_S0</th>\n",
       "      <th>conn_state_S1</th>\n",
       "      <th>conn_state_S3</th>\n",
       "      <th>conn_state_SF</th>\n",
       "      <th>history_C</th>\n",
       "      <th>history_D</th>\n",
       "      <th>history_Dd</th>\n",
       "      <th>history_Other</th>\n",
       "      <th>history_S</th>\n",
       "      <th>history_ShAdDaf</th>\n",
       "      <th>history_ShAdDaft</th>\n",
       "      <th>history_ShAdfDr</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</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>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</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>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</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>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</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>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</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>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>23140</th>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</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>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>23141</th>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</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>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>23142</th>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</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>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</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>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>23143</th>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</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>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</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>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>23144</th>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</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>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</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>0.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>23145 rows × 20 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "       proto_tcp  proto_udp  service_dhcp  service_dns  service_http  \\\n",
       "0            1.0        0.0           0.0          1.0           0.0   \n",
       "1            1.0        0.0           0.0          1.0           0.0   \n",
       "2            1.0        0.0           0.0          1.0           0.0   \n",
       "3            1.0        0.0           0.0          0.0           1.0   \n",
       "4            1.0        0.0           0.0          1.0           0.0   \n",
       "...          ...        ...           ...          ...           ...   \n",
       "23140        1.0        0.0           0.0          0.0           0.0   \n",
       "23141        1.0        0.0           0.0          0.0           0.0   \n",
       "23142        0.0        1.0           1.0          0.0           0.0   \n",
       "23143        0.0        1.0           1.0          0.0           0.0   \n",
       "23144        0.0        1.0           1.0          0.0           0.0   \n",
       "\n",
       "       service_irc  conn_state_OTH  conn_state_RSTR  conn_state_S0  \\\n",
       "0              0.0             0.0              0.0            1.0   \n",
       "1              0.0             0.0              0.0            1.0   \n",
       "2              0.0             0.0              0.0            1.0   \n",
       "3              0.0             0.0              0.0            0.0   \n",
       "4              0.0             0.0              0.0            1.0   \n",
       "...            ...             ...              ...            ...   \n",
       "23140          1.0             0.0              0.0            0.0   \n",
       "23141          1.0             0.0              0.0            0.0   \n",
       "23142          0.0             0.0              0.0            1.0   \n",
       "23143          0.0             0.0              0.0            1.0   \n",
       "23144          0.0             0.0              0.0            1.0   \n",
       "\n",
       "       conn_state_S1  conn_state_S3  conn_state_SF  history_C  history_D  \\\n",
       "0                0.0            0.0            0.0        0.0        0.0   \n",
       "1                0.0            0.0            0.0        0.0        0.0   \n",
       "2                0.0            0.0            0.0        0.0        0.0   \n",
       "3                0.0            0.0            1.0        0.0        0.0   \n",
       "4                0.0            0.0            0.0        0.0        0.0   \n",
       "...              ...            ...            ...        ...        ...   \n",
       "23140            0.0            1.0            0.0        0.0        0.0   \n",
       "23141            0.0            1.0            0.0        0.0        0.0   \n",
       "23142            0.0            0.0            0.0        0.0        1.0   \n",
       "23143            0.0            0.0            0.0        0.0        1.0   \n",
       "23144            0.0            0.0            0.0        0.0        1.0   \n",
       "\n",
       "       history_Dd  history_Other  history_S  history_ShAdDaf  \\\n",
       "0             0.0            0.0        1.0              0.0   \n",
       "1             0.0            0.0        1.0              0.0   \n",
       "2             0.0            0.0        1.0              0.0   \n",
       "3             0.0            1.0        0.0              0.0   \n",
       "4             0.0            0.0        1.0              0.0   \n",
       "...           ...            ...        ...              ...   \n",
       "23140         0.0            0.0        0.0              1.0   \n",
       "23141         0.0            0.0        0.0              1.0   \n",
       "23142         0.0            0.0        0.0              0.0   \n",
       "23143         0.0            0.0        0.0              0.0   \n",
       "23144         0.0            0.0        0.0              0.0   \n",
       "\n",
       "       history_ShAdDaft  history_ShAdfDr  \n",
       "0                   0.0              0.0  \n",
       "1                   0.0              0.0  \n",
       "2                   0.0              0.0  \n",
       "3                   0.0              0.0  \n",
       "4                   0.0              0.0  \n",
       "...                 ...              ...  \n",
       "23140               0.0              0.0  \n",
       "23141               0.0              0.0  \n",
       "23142               0.0              0.0  \n",
       "23143               0.0              0.0  \n",
       "23144               0.0              0.0  \n",
       "\n",
       "[23145 rows x 20 columns]"
      ]
     },
     "execution_count": 98,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Initialize the encoder with its default parameters\n",
    "ohe = OneHotEncoder()\n",
    "\n",
    "# Fit the encoder to categorical features in the dataset\n",
    "encoded_features = ohe.fit_transform(data_df[categorical_features])\n",
    "\n",
    "# Create a dataframe of encoded features\n",
    "encoded_features_df = pd.DataFrame(encoded_features.toarray(), columns=ohe.get_feature_names_out())\n",
    "\n",
    "# Check the results of encoding\n",
    "encoded_features_df"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 99,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Merge encoded features with the dataset and drop original columns\n",
    "data_df = pd.concat([data_df, encoded_features_df], axis=1).drop(categorical_features, axis=1)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 4. STORING PROCESSED DATASET"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Check the processed dataset"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 100,
   "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>id.orig_p</th>\n",
       "      <th>id.resp_p</th>\n",
       "      <th>duration</th>\n",
       "      <th>orig_bytes</th>\n",
       "      <th>resp_bytes</th>\n",
       "      <th>missed_bytes</th>\n",
       "      <th>orig_pkts</th>\n",
       "      <th>orig_ip_bytes</th>\n",
       "      <th>resp_pkts</th>\n",
       "      <th>resp_ip_bytes</th>\n",
       "      <th>label</th>\n",
       "      <th>proto_tcp</th>\n",
       "      <th>proto_udp</th>\n",
       "      <th>service_dhcp</th>\n",
       "      <th>service_dns</th>\n",
       "      <th>service_http</th>\n",
       "      <th>service_irc</th>\n",
       "      <th>conn_state_OTH</th>\n",
       "      <th>conn_state_RSTR</th>\n",
       "      <th>conn_state_S0</th>\n",
       "      <th>conn_state_S1</th>\n",
       "      <th>conn_state_S3</th>\n",
       "      <th>conn_state_SF</th>\n",
       "      <th>history_C</th>\n",
       "      <th>history_D</th>\n",
       "      <th>history_Dd</th>\n",
       "      <th>history_Other</th>\n",
       "      <th>history_S</th>\n",
       "      <th>history_ShAdDaf</th>\n",
       "      <th>history_ShAdDaft</th>\n",
       "      <th>history_ShAdfDr</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0.628686</td>\n",
       "      <td>0.001238</td>\n",
       "      <td>0.620050</td>\n",
       "      <td>0.000000e+00</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000163</td>\n",
       "      <td>2.366458e-06</td>\n",
       "      <td>0.00000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</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>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0.628686</td>\n",
       "      <td>0.001238</td>\n",
       "      <td>0.620022</td>\n",
       "      <td>3.097425e-07</td>\n",
       "      <td>0.000047</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000054</td>\n",
       "      <td>7.888192e-07</td>\n",
       "      <td>0.00000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</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>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0.628686</td>\n",
       "      <td>0.001238</td>\n",
       "      <td>0.620022</td>\n",
       "      <td>3.097425e-07</td>\n",
       "      <td>0.000047</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000054</td>\n",
       "      <td>7.888192e-07</td>\n",
       "      <td>0.00000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</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>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0.628686</td>\n",
       "      <td>0.001238</td>\n",
       "      <td>0.221583</td>\n",
       "      <td>1.972292e-06</td>\n",
       "      <td>0.780758</td>\n",
       "      <td>0.5</td>\n",
       "      <td>0.005097</td>\n",
       "      <td>7.263710e-05</td>\n",
       "      <td>0.08972</td>\n",
       "      <td>0.823184</td>\n",
       "      <td>0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</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>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0.628717</td>\n",
       "      <td>0.001238</td>\n",
       "      <td>0.621946</td>\n",
       "      <td>0.000000e+00</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000163</td>\n",
       "      <td>2.366458e-06</td>\n",
       "      <td>0.00000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</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>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   id.orig_p  id.resp_p  duration    orig_bytes  resp_bytes  missed_bytes  \\\n",
       "0   0.628686   0.001238  0.620050  0.000000e+00    0.000000           0.0   \n",
       "1   0.628686   0.001238  0.620022  3.097425e-07    0.000047           0.0   \n",
       "2   0.628686   0.001238  0.620022  3.097425e-07    0.000047           0.0   \n",
       "3   0.628686   0.001238  0.221583  1.972292e-06    0.780758           0.5   \n",
       "4   0.628717   0.001238  0.621946  0.000000e+00    0.000000           0.0   \n",
       "\n",
       "   orig_pkts  orig_ip_bytes  resp_pkts  resp_ip_bytes  label  proto_tcp  \\\n",
       "0   0.000163   2.366458e-06    0.00000       0.000000      0        1.0   \n",
       "1   0.000054   7.888192e-07    0.00000       0.000000      0        1.0   \n",
       "2   0.000054   7.888192e-07    0.00000       0.000000      0        1.0   \n",
       "3   0.005097   7.263710e-05    0.08972       0.823184      0        1.0   \n",
       "4   0.000163   2.366458e-06    0.00000       0.000000      0        1.0   \n",
       "\n",
       "   proto_udp  service_dhcp  service_dns  service_http  service_irc  \\\n",
       "0        0.0           0.0          1.0           0.0          0.0   \n",
       "1        0.0           0.0          1.0           0.0          0.0   \n",
       "2        0.0           0.0          1.0           0.0          0.0   \n",
       "3        0.0           0.0          0.0           1.0          0.0   \n",
       "4        0.0           0.0          1.0           0.0          0.0   \n",
       "\n",
       "   conn_state_OTH  conn_state_RSTR  conn_state_S0  conn_state_S1  \\\n",
       "0             0.0              0.0            1.0            0.0   \n",
       "1             0.0              0.0            1.0            0.0   \n",
       "2             0.0              0.0            1.0            0.0   \n",
       "3             0.0              0.0            0.0            0.0   \n",
       "4             0.0              0.0            1.0            0.0   \n",
       "\n",
       "   conn_state_S3  conn_state_SF  history_C  history_D  history_Dd  \\\n",
       "0            0.0            0.0        0.0        0.0         0.0   \n",
       "1            0.0            0.0        0.0        0.0         0.0   \n",
       "2            0.0            0.0        0.0        0.0         0.0   \n",
       "3            0.0            1.0        0.0        0.0         0.0   \n",
       "4            0.0            0.0        0.0        0.0         0.0   \n",
       "\n",
       "   history_Other  history_S  history_ShAdDaf  history_ShAdDaft  \\\n",
       "0            0.0        1.0              0.0               0.0   \n",
       "1            0.0        1.0              0.0               0.0   \n",
       "2            0.0        1.0              0.0               0.0   \n",
       "3            1.0        0.0              0.0               0.0   \n",
       "4            0.0        1.0              0.0               0.0   \n",
       "\n",
       "   history_ShAdfDr  \n",
       "0              0.0  \n",
       "1              0.0  \n",
       "2              0.0  \n",
       "3              0.0  \n",
       "4              0.0  "
      ]
     },
     "execution_count": 100,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Check the final version of the dataset\n",
    "data_df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 101,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<class 'pandas.core.frame.DataFrame'>\n",
      "Int64Index: 23145 entries, 0 to 23144\n",
      "Data columns (total 31 columns):\n",
      " #   Column            Non-Null Count  Dtype  \n",
      "---  ------            --------------  -----  \n",
      " 0   id.orig_p         23145 non-null  float64\n",
      " 1   id.resp_p         23145 non-null  float64\n",
      " 2   duration          23145 non-null  float64\n",
      " 3   orig_bytes        23145 non-null  float64\n",
      " 4   resp_bytes        23145 non-null  float64\n",
      " 5   missed_bytes      23145 non-null  float64\n",
      " 6   orig_pkts         23145 non-null  float64\n",
      " 7   orig_ip_bytes     23145 non-null  float64\n",
      " 8   resp_pkts         23145 non-null  float64\n",
      " 9   resp_ip_bytes     23145 non-null  float64\n",
      " 10  label             23145 non-null  int32  \n",
      " 11  proto_tcp         23145 non-null  float64\n",
      " 12  proto_udp         23145 non-null  float64\n",
      " 13  service_dhcp      23145 non-null  float64\n",
      " 14  service_dns       23145 non-null  float64\n",
      " 15  service_http      23145 non-null  float64\n",
      " 16  service_irc       23145 non-null  float64\n",
      " 17  conn_state_OTH    23145 non-null  float64\n",
      " 18  conn_state_RSTR   23145 non-null  float64\n",
      " 19  conn_state_S0     23145 non-null  float64\n",
      " 20  conn_state_S1     23145 non-null  float64\n",
      " 21  conn_state_S3     23145 non-null  float64\n",
      " 22  conn_state_SF     23145 non-null  float64\n",
      " 23  history_C         23145 non-null  float64\n",
      " 24  history_D         23145 non-null  float64\n",
      " 25  history_Dd        23145 non-null  float64\n",
      " 26  history_Other     23145 non-null  float64\n",
      " 27  history_S         23145 non-null  float64\n",
      " 28  history_ShAdDaf   23145 non-null  float64\n",
      " 29  history_ShAdDaft  23145 non-null  float64\n",
      " 30  history_ShAdfDr   23145 non-null  float64\n",
      "dtypes: float64(30), int32(1)\n",
      "memory usage: 5.6 MB\n"
     ]
    }
   ],
   "source": [
    "# Check not null values and attribute types\n",
    "data_df.info()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Store processed dataset to a csv file"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 102,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Set file location and name\n",
    "data_file_location = \"..\\\\data\\\\processed\\\\\"\n",
    "data_file_name = \"conn.log.labeled_processed\"\n",
    "data_file_ext = \".csv\"\n",
    "\n",
    "# Store processed dataset to a csv file\n",
    "data_df.to_csv(data_file_location + data_file_name + data_file_ext)"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "base",
   "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"
  },
  "orig_nbformat": 4
 },
 "nbformat": 4,
 "nbformat_minor": 2
}