VPN_12.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "id": "56db7f42",
   "metadata": {},
   "source": [
    "## Prepare Dataset"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "8598fd5e-a7d0-4b9d-bc9a-fa413c9fde19",
   "metadata": {},
   "source": [
    "### Flow count for VPN and non-VPN"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "771f0e5f-bbf4-4226-86c5-d4ac170fd04a",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "label\n",
      "NoVPN    43597\n",
      "VPN       8350\n",
      "Name: count, dtype: int64\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 800x600 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "import pandas as pd\n",
    "import matplotlib.pyplot as plt\n",
    "\n",
    "df = pd.read_csv('./VNAT/WF/flows_extra_labelled_12.csv',low_memory=False)\n",
    "\n",
    "label_counts = df['label'].value_counts()\n",
    "print(label_counts)\n",
    "\n",
    "plt.figure(figsize=(8, 6))  \n",
    "label_counts.plot(kind='bar', color=['skyblue', 'red'])  \n",
    "\n",
    "plt.title('Count of Labels in CSV', fontsize=16)\n",
    "plt.xlabel('Label', fontsize=14)\n",
    "plt.ylabel('Count', fontsize=14)\n",
    "\n",
    "plt.show()\n"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "97bf462c-e9b0-4f61-a336-9529b0eb9837",
   "metadata": {},
   "source": [
    "### Create a dataset for Wavelet Features"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "3d29e0a5-e2a3-4794-8ec4-27b03d525153",
   "metadata": {},
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "import numpy as np\n",
    "\n",
    "df1 = pd.read_csv('./VNAT/WF/flows_extra_labelled_12.csv',low_memory=False)\n",
    "\n",
    "\n",
    "columns_to_select = [\n",
    "                     #select Wavelet features\n",
    "                     #forward\n",
    "                    'udps.f_p_k_0','udps.f_p_k_1','udps.f_p_k_2','udps.f_p_k_3','udps.f_p_k_4','udps.f_p_k_5','udps.f_p_k_6',\n",
    "                     'udps.f_p_k_7','udps.f_p_k_8','udps.f_p_k_9','udps.f_p_k_10','udps.f_p_k_11','udps.f_p_k_12',\n",
    "\n",
    "                     'udps.f_u_k_0','udps.f_u_k_1','udps.f_u_k_2','udps.f_u_k_3','udps.f_u_k_4','udps.f_u_k_5','udps.f_u_k_6',\n",
    "                     'udps.f_u_k_7','udps.f_u_k_8','udps.f_u_k_9','udps.f_u_k_10','udps.f_u_k_11','udps.f_u_k_12',\n",
    "                     \n",
    "                     'udps.f_sigma_k_0','udps.f_sigma_k_1','udps.f_sigma_k_2','udps.f_sigma_k_3','udps.f_sigma_k_4','udps.f_sigma_k_5',\n",
    "           'udps.f_sigma_k_6','udps.f_sigma_k_7','udps.f_sigma_k_8','udps.f_sigma_k_9','udps.f_sigma_k_10','udps.f_sigma_k_11','udps.f_sigma_k_12',\n",
    "\n",
    "                     'udps.f_S_k_0','udps.f_S_k_1','udps.f_S_k_2','udps.f_S_k_3','udps.f_S_k_4','udps.f_S_k_5','udps.f_S_k_6','udps.f_S_k_7',\n",
    "                     'udps.f_S_k_8','udps.f_S_k_9','udps.f_S_k_10','udps.f_S_k_11','udps.f_S_k_12',\n",
    "                      #backward\n",
    "                     'udps.b_p_k_0','udps.b_p_k_1','udps.b_p_k_2','udps.b_p_k_3','udps.b_p_k_4','udps.b_p_k_5','udps.b_p_k_6',\n",
    "                     'udps.b_p_k_7','udps.b_p_k_8','udps.b_p_k_9','udps.b_p_k_10','udps.b_p_k_11','udps.b_p_k_12',\n",
    "\n",
    "                     'udps.b_u_k_0','udps.b_u_k_1','udps.b_u_k_2','udps.b_u_k_3','udps.b_u_k_4','udps.b_u_k_5','udps.b_u_k_6',\n",
    "                     'udps.b_u_k_7','udps.b_u_k_8','udps.b_u_k_9','udps.b_u_k_10','udps.b_u_k_11','udps.b_u_k_12',\n",
    "                     \n",
    "                     'udps.b_sigma_k_0','udps.b_sigma_k_1','udps.b_sigma_k_2','udps.b_sigma_k_3','udps.b_sigma_k_4','udps.b_sigma_k_5',\n",
    "           'udps.b_sigma_k_6','udps.b_sigma_k_7','udps.b_sigma_k_8','udps.b_sigma_k_9','udps.b_sigma_k_10','udps.b_sigma_k_11','udps.b_sigma_k_12',\n",
    "\n",
    "                     'udps.b_S_k_0','udps.b_S_k_1','udps.b_S_k_2','udps.b_S_k_3','udps.b_S_k_4','udps.b_S_k_5','udps.b_S_k_6','udps.b_S_k_7',\n",
    "                     'udps.b_S_k_8','udps.b_S_k_9','udps.b_S_k_10','udps.b_S_k_11','udps.b_S_k_12'\n",
    "                   , 'label','category']\n",
    "df_WNF = df1[columns_to_select]                  \n",
    "\n",
    "df_WNF.to_csv('./VNAT/WF/flows_extra_labelled_WF12.csv', index=False)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "e14e4aad-250e-44d4-b5f4-3bfa61ac91c4",
   "metadata": {},
   "source": [
    "### Rndom Forest"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "409c26c3-2718-4ace-bc59-6542edc3b059",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train Accuracy: 0.999975936665303\n",
      "Test Accuracy: 0.9932627526467758\n",
      "\n",
      "Confusion Matrix:\n",
      "[[0.99713336 0.00286664]\n",
      " [0.02696225 0.97303775]]\n",
      "\n",
      "Classification Report:\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "       NoVPN       0.99      1.00      1.00      8721\n",
      "         VPN       0.98      0.97      0.98      1669\n",
      "\n",
      "    accuracy                           0.99     10390\n",
      "   macro avg       0.99      0.99      0.99     10390\n",
      "weighted avg       0.99      0.99      0.99     10390\n",
      "\n",
      "\n",
      "Detailed Misclassification Summary (per Category):\n",
      "            category true_label predicted_label  count\n",
      "0               Chat      NoVPN             VPN     17\n",
      "1               Chat        VPN           NoVPN     29\n",
      "2  Command & Control      NoVPN             VPN      2\n",
      "3  Command & Control        VPN           NoVPN      4\n",
      "4      File Transfer        VPN           NoVPN      8\n",
      "5          Streaming      NoVPN             VPN      6\n",
      "6          Streaming        VPN           NoVPN      4\n",
      "Confusion matrix saved at: output_results\\confusion_matrix_RF12.png\n",
      "Accuracy and classification report saved at: output_results\\RF12.txt\n",
      "Detailed misclassification summary saved at: output_results\\RF12.csv\n"
     ]
    }
   ],
   "source": [
    "import pandas as pd\n",
    "import numpy as np\n",
    "import os\n",
    "import matplotlib.pyplot as plt\n",
    "import seaborn as sns\n",
    "from sklearn.model_selection import train_test_split\n",
    "from sklearn.ensemble import RandomForestClassifier\n",
    "from sklearn.metrics import classification_report, confusion_matrix, accuracy_score\n",
    "\n",
    "output_dir = \"output_results\"\n",
    "os.makedirs(output_dir, exist_ok=True)\n",
    "\n",
    "file_path = './VNAT/WF/flows_extra_labelled_WF12.csv'  \n",
    "df = pd.read_csv(file_path)\n",
    "\n",
    "df['label_encoded'] = df['label'].map({'NoVPN': 0, 'VPN': 1})\n",
    "\n",
    "category = df['category']\n",
    "\n",
    "X = df.drop(['label', 'label_encoded', 'category'], axis=1)\n",
    "y = df['label_encoded']\n",
    "\n",
    "X_train, X_test, y_train, y_test, category_train, category_test = train_test_split(\n",
    "    X, y, category, test_size=0.2, random_state=42\n",
    ")\n",
    "\n",
    "clf = RandomForestClassifier(random_state=42)\n",
    "clf.fit(X_train, y_train)\n",
    "\n",
    "y_pred = clf.predict(X_test)\n",
    "\n",
    "train_accuracy = accuracy_score(y_train, clf.predict(X_train))\n",
    "test_accuracy = accuracy_score(y_test, y_pred)\n",
    "\n",
    "cm = confusion_matrix(y_test, y_pred, normalize='true')\n",
    "\n",
    "print(\"Train Accuracy:\", train_accuracy)\n",
    "print(\"Test Accuracy:\", test_accuracy)\n",
    "print(\"\\nConfusion Matrix:\")\n",
    "print(cm)\n",
    "print(\"\\nClassification Report:\")\n",
    "print(classification_report(y_test, y_pred, target_names=['NoVPN', 'VPN']))\n",
    "\n",
    "plt.figure(figsize=(6, 5))\n",
    "sns.heatmap(cm, annot=True, fmt=\".2f\", cmap=\"Blues\", xticklabels=['non-VPN', 'VPN'], \n",
    "yticklabels=['non-VPN', 'VPN'], annot_kws={\"size\": 16}) \n",
    "\n",
    "plt.xlabel(\"Predicted Label\", fontsize=20)\n",
    "plt.ylabel(\"True Label\", fontsize=20)\n",
    "plt.title(\"\", fontsize=20)\n",
    "plt.xticks(fontsize=20)\n",
    "plt.yticks(fontsize=20)\n",
    "plt.tight_layout()\n",
    "\n",
    "conf_matrix_path = os.path.join(output_dir, \"confusion_matrix_RF12.png\")\n",
    "plt.savefig(conf_matrix_path, dpi=300, bbox_inches='tight')\n",
    "plt.close()\n",
    "\n",
    "accuracy_report_path = os.path.join(output_dir, \"RF12.txt\")\n",
    "with open(accuracy_report_path, \"w\") as f:\n",
    "    f.write(f\"Train Accuracy: {train_accuracy:.4f}\\n\")\n",
    "    \n",
    "    f.write(f\"Test Accuracy: {test_accuracy:.4f}\\n\\n\")\n",
    "    f.write(\"Classification Report:\\n\")\n",
    "    f.write(classification_report(y_test, y_pred, target_names=['NoVPN', 'VPN']))\n",
    "\n",
    "misclassified_mask = (y_test != y_pred)\n",
    "misclassified_samples = X_test.loc[misclassified_mask].copy()\n",
    "\n",
    "misclassified_samples['true_label'] = y_test.loc[misclassified_mask]\n",
    "misclassified_samples['predicted_label'] = y_pred[misclassified_mask]\n",
    "misclassified_samples['category'] = category_test.loc[misclassified_mask]\n",
    "\n",
    "label_map = {0: 'NoVPN', 1: 'VPN'}\n",
    "misclassified_samples['true_label'] = misclassified_samples['true_label'].map(label_map)\n",
    "misclassified_samples['predicted_label'] = misclassified_samples['predicted_label'].map(label_map)\n",
    "\n",
    "misclassification_detail = misclassified_samples.groupby(['category', 'true_label', 'predicted_label']).size().reset_index(name='count')\n",
    "\n",
    "print(\"\\nDetailed Misclassification Summary (per Category):\")\n",
    "print(misclassification_detail)\n",
    "\n",
    "misclassification_detail_path = os.path.join(output_dir, \"RF12.csv\")\n",
    "misclassification_detail.to_csv(misclassification_detail_path, index=False)\n",
    "\n",
    "print(\"Confusion matrix saved at:\", conf_matrix_path)\n",
    "print(\"Accuracy and classification report saved at:\", accuracy_report_path)\n",
    "print(\"Detailed misclassification summary saved at:\", misclassification_detail_path)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "id": "024cf7b6-2531-49e4-865c-fee06c3051db",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 1200x600 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "#  feature importances\n",
    "importances = clf.feature_importances_\n",
    "indices = np.argsort(importances)[::-1]\n",
    "\n",
    "plt.figure(figsize=(12, 6))\n",
    "plt.title(\"Feature importances\")\n",
    "plt.bar(range(X_train.shape[1]), importances[indices],\n",
    "        color=\"r\", align=\"center\")\n",
    "plt.xticks(range(X_train.shape[1]), X_train.columns[indices], rotation=90)\n",
    "plt.xlim([-1, X_train.shape[1]+1])\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "1d89b9de-19ce-4293-b3ae-afabe3a67cb4",
   "metadata": {},
   "source": [
    "### SVM"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "e097c028-e83f-4496-83a8-d607bc93296e",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train Accuracy: 0.8993911976321679\n",
      "Test Accuracy: 0.8962463907603465\n",
      "\n",
      "Confusion Matrix:\n",
      "[[0.93165921 0.06834079]\n",
      " [0.28879569 0.71120431]]\n",
      "\n",
      "Classification Report:\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "       NoVPN       0.94      0.93      0.94      8721\n",
      "         VPN       0.67      0.71      0.69      1669\n",
      "\n",
      "    accuracy                           0.90     10390\n",
      "   macro avg       0.80      0.82      0.81     10390\n",
      "weighted avg       0.90      0.90      0.90     10390\n",
      "\n",
      "Accuracy and classification report saved.\n",
      "\n",
      "Detailed Misclassification Summary (per Category):\n",
      "            category true_label predicted_label  count\n",
      "0               Chat      NoVPN             VPN    459\n",
      "1               Chat        VPN           NoVPN    311\n",
      "2  Command & Control      NoVPN             VPN     13\n",
      "3  Command & Control        VPN           NoVPN     51\n",
      "4      File Transfer      NoVPN             VPN     27\n",
      "5      File Transfer        VPN           NoVPN     31\n",
      "6          Streaming      NoVPN             VPN     50\n",
      "7          Streaming        VPN           NoVPN     27\n",
      "8               VoIP      NoVPN             VPN     47\n",
      "9               VoIP        VPN           NoVPN     62\n",
      "Confusion matrix saved at: output_results\\confusion_matrix_SVM12.png\n",
      "Accuracy and classification report saved at: output_results\\SVM12.txt\n",
      "Detailed misclassification summary saved at: output_results\\SVM12.csv\n"
     ]
    }
   ],
   "source": [
    "import pandas as pd\n",
    "import numpy as np\n",
    "import os\n",
    "import matplotlib.pyplot as plt\n",
    "import seaborn as sns\n",
    "from sklearn.model_selection import train_test_split\n",
    "from sklearn.svm import LinearSVC\n",
    "from sklearn.metrics import classification_report, confusion_matrix, accuracy_score\n",
    "from sklearn.preprocessing import StandardScaler\n",
    "\n",
    "output_dir = \"output_results\"\n",
    "os.makedirs(output_dir, exist_ok=True)\n",
    "\n",
    "file_path = './VNAT/WF/flows_extra_labelled_WF12.csv'  # Update path if needed\n",
    "df = pd.read_csv(file_path)\n",
    "\n",
    "df['label_encoded'] = df['label'].map({'NoVPN': 0, 'VPN': 1})\n",
    "\n",
    "category = df['category']\n",
    "\n",
    "X = df.drop(['label', 'label_encoded', 'category'], axis=1)\n",
    "y = df['label_encoded']\n",
    "\n",
    "\n",
    "X_train, X_test, y_train, y_test, category_train, category_test = train_test_split(\n",
    "    X, y, category, test_size=0.2, random_state=42\n",
    ")\n",
    "\n",
    "scaler = StandardScaler()\n",
    "X_train = pd.DataFrame(scaler.fit_transform(X_train), columns=X.columns, index=X_train.index)\n",
    "X_test = pd.DataFrame(scaler.transform(X_test), columns=X.columns, index=X_test.index)\n",
    "\n",
    "\n",
    "clf = LinearSVC(random_state=42, max_iter=10000)\n",
    "clf.fit(X_train, y_train)\n",
    "\n",
    "y_pred = clf.predict(X_test)\n",
    "\n",
    "train_accuracy = accuracy_score(y_train, clf.predict(X_train))\n",
    "test_accuracy = accuracy_score(y_test, y_pred)\n",
    "\n",
    "cm = confusion_matrix(y_test, y_pred, normalize='true')\n",
    "\n",
    "print(\"Train Accuracy:\", train_accuracy)\n",
    "print(\"Test Accuracy:\", test_accuracy)\n",
    "print(\"\\nConfusion Matrix:\")\n",
    "print(cm)\n",
    "print(\"\\nClassification Report:\")\n",
    "print(classification_report(y_test, y_pred, target_names=['NoVPN', 'VPN']))\n",
    "\n",
    "plt.figure(figsize=(6, 5))\n",
    "sns.heatmap(cm, annot=True, fmt=\".2f\", cmap=\"Blues\", xticklabels=['NoVPN', 'VPN'], \n",
    "yticklabels=['NoVPN', 'VPN'], annot_kws={\"size\": 16})  # Enlarged numbers inside the matrix\n",
    "\n",
    "plt.xlabel(\"Predicted Label\", fontsize=20)\n",
    "plt.ylabel(\"True Label\", fontsize=20)\n",
    "plt.title(\"\", fontsize=20)\n",
    "plt.xticks(fontsize=20)\n",
    "plt.yticks(fontsize=20)\n",
    "plt.tight_layout()\n",
    "\n",
    "conf_matrix_path = os.path.join(output_dir, \"confusion_matrix_SVM12.png\")\n",
    "plt.savefig(conf_matrix_path, dpi=300, bbox_inches='tight')\n",
    "plt.close()\n",
    "\n",
    "accuracy_report_path = os.path.join(output_dir, \"SVM12.txt\")\n",
    "with open(accuracy_report_path, \"w\") as f:\n",
    "    f.write(f\"Train Accuracy: {train_accuracy:.4f}\\n\")\n",
    "    f.write(f\"Test Accuracy: {test_accuracy:.4f}\\n\\n\")\n",
    "    f.write(\"Classification Report:\\n\")\n",
    "    f.write(classification_report(y_test, y_pred, target_names=['NoVPN', 'VPN']))\n",
    "print(\"Accuracy and classification report saved.\")\n",
    "\n",
    "misclassified_mask = (y_test != y_pred)\n",
    "misclassified_samples = X_test.loc[misclassified_mask].copy()\n",
    "\n",
    "misclassified_samples['true_label'] = y_test.loc[misclassified_mask]\n",
    "misclassified_samples['predicted_label'] = y_pred[misclassified_mask]\n",
    "misclassified_samples['category'] = category_test.loc[misclassified_mask]\n",
    "\n",
    "label_map = {0: 'NoVPN', 1: 'VPN'}\n",
    "misclassified_samples['true_label'] = misclassified_samples['true_label'].map(label_map)\n",
    "misclassified_samples['predicted_label'] = misclassified_samples['predicted_label'].map(label_map)\n",
    "\n",
    "misclassification_detail = misclassified_samples.groupby(['category', 'true_label', 'predicted_label']).size().reset_index(name='count')\n",
    "\n",
    "print(\"\\nDetailed Misclassification Summary (per Category):\")\n",
    "print(misclassification_detail)\n",
    "\n",
    "misclassification_detail_path = os.path.join(output_dir, \"SVM12.csv\")\n",
    "misclassification_detail.to_csv(misclassification_detail_path, index=False)\n",
    "\n",
    "print(\"Confusion matrix saved at:\", conf_matrix_path)\n",
    "print(\"Accuracy and classification report saved at:\", accuracy_report_path)\n",
    "print(\"Detailed misclassification summary saved at:\", misclassification_detail_path)\n"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "6b8f5108-3b9a-448e-bc32-a5d7464e2913",
   "metadata": {},
   "source": [
    "### Neural Network"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "id": "bbd06971-d739-43ef-bca3-3599021f5404",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/20\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\yasah\\anaconda3\\envs\\py311\\Lib\\site-packages\\keras\\src\\layers\\core\\dense.py:87: UserWarning: Do not pass an `input_shape`/`input_dim` argument to a layer. When using Sequential models, prefer using an `Input(shape)` object as the first layer in the model instead.\n",
      "  super().__init__(activity_regularizer=activity_regularizer, **kwargs)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\u001b[1m1299/1299\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m7s\u001b[0m 3ms/step - accuracy: 0.8891 - loss: 0.2258 - val_accuracy: 0.9556 - val_loss: 0.1273\n",
      "Epoch 2/20\n",
      "\u001b[1m1299/1299\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m4s\u001b[0m 3ms/step - accuracy: 0.9560 - loss: 0.1152 - val_accuracy: 0.9662 - val_loss: 0.0880\n",
      "Epoch 3/20\n",
      "\u001b[1m1299/1299\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m4s\u001b[0m 3ms/step - accuracy: 0.9680 - loss: 0.0855 - val_accuracy: 0.9680 - val_loss: 0.0817\n",
      "Epoch 4/20\n",
      "\u001b[1m1299/1299\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m4s\u001b[0m 3ms/step - accuracy: 0.9740 - loss: 0.0705 - val_accuracy: 0.9756 - val_loss: 0.0651\n",
      "Epoch 5/20\n",
      "\u001b[1m1299/1299\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m4s\u001b[0m 3ms/step - accuracy: 0.9765 - loss: 0.0625 - val_accuracy: 0.9734 - val_loss: 0.0669\n",
      "Epoch 6/20\n",
      "\u001b[1m1299/1299\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m6s\u001b[0m 4ms/step - accuracy: 0.9801 - loss: 0.0539 - val_accuracy: 0.9794 - val_loss: 0.0601\n",
      "Epoch 7/20\n",
      "\u001b[1m1299/1299\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m5s\u001b[0m 4ms/step - accuracy: 0.9791 - loss: 0.0550 - val_accuracy: 0.9797 - val_loss: 0.0548\n",
      "Epoch 8/20\n",
      "\u001b[1m1299/1299\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m3s\u001b[0m 3ms/step - accuracy: 0.9821 - loss: 0.0500 - val_accuracy: 0.9789 - val_loss: 0.0581\n",
      "Epoch 9/20\n",
      "\u001b[1m1299/1299\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m4s\u001b[0m 3ms/step - accuracy: 0.9832 - loss: 0.0457 - val_accuracy: 0.9833 - val_loss: 0.0483\n",
      "Epoch 10/20\n",
      "\u001b[1m1299/1299\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m3s\u001b[0m 3ms/step - accuracy: 0.9850 - loss: 0.0414 - val_accuracy: 0.9822 - val_loss: 0.0482\n",
      "Epoch 11/20\n",
      "\u001b[1m1299/1299\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m3s\u001b[0m 2ms/step - accuracy: 0.9850 - loss: 0.0406 - val_accuracy: 0.9833 - val_loss: 0.0450\n",
      "Epoch 12/20\n",
      "\u001b[1m1299/1299\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m3s\u001b[0m 2ms/step - accuracy: 0.9868 - loss: 0.0380 - val_accuracy: 0.9828 - val_loss: 0.0455\n",
      "Epoch 13/20\n",
      "\u001b[1m1299/1299\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m5s\u001b[0m 2ms/step - accuracy: 0.9865 - loss: 0.0368 - val_accuracy: 0.9822 - val_loss: 0.0471\n",
      "Epoch 14/20\n",
      "\u001b[1m1299/1299\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m4s\u001b[0m 3ms/step - accuracy: 0.9869 - loss: 0.0365 - val_accuracy: 0.9852 - val_loss: 0.0462\n",
      "Epoch 15/20\n",
      "\u001b[1m1299/1299\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m4s\u001b[0m 3ms/step - accuracy: 0.9873 - loss: 0.0364 - val_accuracy: 0.9850 - val_loss: 0.0447\n",
      "Epoch 16/20\n",
      "\u001b[1m1299/1299\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m4s\u001b[0m 3ms/step - accuracy: 0.9889 - loss: 0.0322 - val_accuracy: 0.9831 - val_loss: 0.0475\n",
      "Epoch 17/20\n",
      "\u001b[1m1299/1299\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m4s\u001b[0m 3ms/step - accuracy: 0.9885 - loss: 0.0342 - val_accuracy: 0.9819 - val_loss: 0.0491\n",
      "Epoch 18/20\n",
      "\u001b[1m1299/1299\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m4s\u001b[0m 3ms/step - accuracy: 0.9887 - loss: 0.0328 - val_accuracy: 0.9835 - val_loss: 0.0447\n",
      "Epoch 19/20\n",
      "\u001b[1m1299/1299\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m4s\u001b[0m 3ms/step - accuracy: 0.9895 - loss: 0.0288 - val_accuracy: 0.9850 - val_loss: 0.0467\n",
      "Epoch 20/20\n",
      "\u001b[1m1299/1299\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m4s\u001b[0m 3ms/step - accuracy: 0.9902 - loss: 0.0310 - val_accuracy: 0.9848 - val_loss: 0.0463\n",
      "\u001b[1m325/325\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m1s\u001b[0m 2ms/step\n",
      "Train Accuracy: 0.99095219373703\n",
      "Test Accuracy: 0.9847930702598653\n",
      "\n",
      "Confusion Matrix:\n",
      "[[0.98807476 0.01192524]\n",
      " [0.0323547  0.9676453 ]]\n",
      "\n",
      "Classification Report:\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "       NoVPN       0.99      0.99      0.99      8721\n",
      "         VPN       0.94      0.97      0.95      1669\n",
      "\n",
      "    accuracy                           0.98     10390\n",
      "   macro avg       0.97      0.98      0.97     10390\n",
      "weighted avg       0.99      0.98      0.98     10390\n",
      "\n",
      "\n",
      "Detailed Misclassification Summary (per Category):\n",
      "            category true_label predicted_label  count\n",
      "0               Chat      NoVPN             VPN     65\n",
      "1               Chat        VPN           NoVPN     27\n",
      "2  Command & Control      NoVPN             VPN      8\n",
      "3  Command & Control        VPN           NoVPN      5\n",
      "4      File Transfer      NoVPN             VPN      7\n",
      "5      File Transfer        VPN           NoVPN     15\n",
      "6          Streaming      NoVPN             VPN     18\n",
      "7          Streaming        VPN           NoVPN      5\n",
      "8               VoIP      NoVPN             VPN      6\n",
      "9               VoIP        VPN           NoVPN      2\n",
      "Confusion matrix saved at: output_results\\confusion_matrix_NN12.png\n",
      "Accuracy and classification report saved at: output_results\\NN12.txt\n",
      "Detailed misclassification summary saved at: output_results\\NN12.csv\n"
     ]
    }
   ],
   "source": [
    "import pandas as pd\n",
    "import numpy as np\n",
    "import os\n",
    "import matplotlib.pyplot as plt\n",
    "import seaborn as sns\n",
    "from sklearn.model_selection import train_test_split\n",
    "from sklearn.preprocessing import StandardScaler\n",
    "from sklearn.metrics import classification_report, confusion_matrix, accuracy_score\n",
    "import tensorflow as tf\n",
    "from tensorflow import keras\n",
    "\n",
    "output_dir = \"output_results\"\n",
    "os.makedirs(output_dir, exist_ok=True)\n",
    "\n",
    "file_path = './VNAT/WF/flows_extra_labelled_WF12.csv'  \n",
    "df = pd.read_csv(file_path)\n",
    "\n",
    "df['label_encoded'] = df['label'].map({'NoVPN': 0, 'VPN': 1})\n",
    "\n",
    "category = df['category']\n",
    "\n",
    "X = df.drop(['label', 'label_encoded', 'category'], axis=1)\n",
    "y = df['label_encoded']\n",
    "\n",
    "X_train, X_test, y_train, y_test, category_train, category_test = train_test_split(\n",
    "    X, y, category, test_size=0.2, random_state=42\n",
    ")\n",
    "\n",
    "scaler = StandardScaler()\n",
    "X_train = scaler.fit_transform(X_train)\n",
    "X_test = scaler.transform(X_test)\n",
    "\n",
    "model = keras.Sequential([\n",
    "    keras.layers.Dense(64, activation='relu', input_shape=(X_train.shape[1],)),\n",
    "    keras.layers.Dense(32, activation='relu'),\n",
    "    keras.layers.Dense(1, activation='sigmoid')\n",
    "])\n",
    "\n",
    "model.compile(optimizer='adam', loss='binary_crossentropy', metrics=['accuracy'])\n",
    "\n",
    "model.fit(X_train, y_train, epochs=20, batch_size=32, validation_data=(X_test, y_test), verbose=1)\n",
    "\n",
    "y_pred_probs = model.predict(X_test)\n",
    "y_pred = (y_pred_probs > 0.5).astype(int).flatten()\n",
    "\n",
    "train_accuracy = model.evaluate(X_train, y_train, verbose=0)[1]\n",
    "test_accuracy = accuracy_score(y_test, y_pred)\n",
    "\n",
    "cm = confusion_matrix(y_test, y_pred, normalize='true')\n",
    "\n",
    "print(\"Train Accuracy:\", train_accuracy)\n",
    "print(\"Test Accuracy:\", test_accuracy)\n",
    "print(\"\\nConfusion Matrix:\")\n",
    "print(cm)\n",
    "print(\"\\nClassification Report:\")\n",
    "print(classification_report(y_test, y_pred, target_names=['NoVPN', 'VPN']))\n",
    "\n",
    "plt.figure(figsize=(6, 5))\n",
    "sns.heatmap(cm, annot=True, fmt=\".2f\", cmap=\"Blues\", xticklabels=['NoVPN', 'VPN'], \n",
    "yticklabels=['NoVPN', 'VPN'], annot_kws={\"size\": 16})  # Enlarged numbers inside the matrix\n",
    "\n",
    "plt.xlabel(\"Predicted Label\", fontsize=20)\n",
    "plt.ylabel(\"True Label\", fontsize=20)\n",
    "plt.title(\"\", fontsize=20)\n",
    "plt.xticks(fontsize=20)\n",
    "plt.yticks(fontsize=20)\n",
    "plt.tight_layout()\n",
    "\n",
    "conf_matrix_path = os.path.join(output_dir, \"confusion_matrix_NN12.png\")\n",
    "plt.savefig(conf_matrix_path, dpi=300, bbox_inches='tight')\n",
    "plt.close()\n",
    "\n",
    "accuracy_report_path = os.path.join(output_dir, \"NN12.txt\")\n",
    "with open(accuracy_report_path, \"w\") as f:\n",
    "    f.write(f\"Train Accuracy: {train_accuracy:.4f}\\n\")\n",
    "    f.write(f\"Test Accuracy: {test_accuracy:.4f}\\n\\n\")\n",
    "    f.write(\"Classification Report:\\n\")\n",
    "    f.write(classification_report(y_test, y_pred, target_names=['NoVPN', 'VPN']))\n",
    "\n",
    "misclassified_mask = (y_test != y_pred)\n",
    "misclassified_samples = X_test[misclassified_mask]\n",
    "\n",
    "misclassified_samples = pd.DataFrame(misclassified_samples, columns=X.columns)\n",
    "misclassified_samples['true_label'] = y_test.loc[misclassified_mask].values\n",
    "misclassified_samples['predicted_label'] = y_pred[misclassified_mask]\n",
    "misclassified_samples['category'] = category_test.loc[misclassified_mask].values\n",
    "\n",
    "label_map = {0: 'NoVPN', 1: 'VPN'}\n",
    "misclassified_samples['true_label'] = misclassified_samples['true_label'].map(label_map)\n",
    "misclassified_samples['predicted_label'] = misclassified_samples['predicted_label'].map(label_map)\n",
    "\n",
    "misclassification_detail = misclassified_samples.groupby(['category', 'true_label', 'predicted_label']).size().reset_index(name='count')\n",
    "\n",
    "print(\"\\nDetailed Misclassification Summary (per Category):\")\n",
    "print(misclassification_detail)\n",
    "\n",
    "misclassification_detail_path = os.path.join(output_dir, \"NN12.csv\")\n",
    "misclassification_detail.to_csv(misclassification_detail_path, index=False)\n",
    "\n",
    "print(\"Confusion matrix saved at:\", conf_matrix_path)\n",
    "print(\"Accuracy and classification report saved at:\", accuracy_report_path)\n",
    "print(\"Detailed misclassification summary saved at:\", misclassification_detail_path)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "159bf0ae-4567-4f51-ab97-df217f5e0dc9",
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3 (ipykernel)",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.11.10"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}