Upload example/example.ipynb with huggingface_hub

example/example.ipynb

@@ -0,0 +1,479 @@
+{
+ "cells": [
+  {
+   "cell_type": "markdown",
+   "id": "fb6a6862-9ab9-47c7-b5da-0bc772897129",
+   "metadata": {},
+   "source": [
+    "# Training a ML model using CICIoT2023\n",
+    "\n",
+    "This notebook shows how a LogisticRegression model can be trained using the CICIoT2023 csv files."
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 1,
+   "id": "40f7c50d-b0ae-4f19-9398-1435ba7a851d",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "import pandas as pd\n",
+    "import numpy as np\n",
+    "import os\n",
+    "from tqdm import tqdm\n",
+    "import warnings\n",
+    "warnings.filterwarnings('ignore')\n",
+    "from sklearn.linear_model import LogisticRegression"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 2,
+   "id": "5c40b5d2-727b-4f37-a480-9d46304eb541",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "DATASET_DIRECTORY = '../CICIoT2023/'"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "3ec1f2b2-92b3-4622-895b-6ac5126f30b4",
+   "metadata": {},
+   "source": [
+    "### Importing Dataset"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 3,
+   "id": "6854f877-5524-46ba-b7ca-5d6040015f44",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "df_sets = [k for k in os.listdir(DATASET_DIRECTORY) if k.endswith('.csv')]\n",
+    "df_sets.sort()\n",
+    "training_sets = df_sets[:int(len(df_sets)*.8)]\n",
+    "test_sets = df_sets[int(len(df_sets)*.8):]"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 4,
+   "id": "0433838d-ca57-4dd8-b41c-ad2ee3df61c4",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "X_columns = [\n",
+    "    'flow_duration', 'Header_Length', 'Protocol Type', 'Duration',\n",
+    "       'Rate', 'Srate', 'Drate', 'fin_flag_number', 'syn_flag_number',\n",
+    "       'rst_flag_number', 'psh_flag_number', 'ack_flag_number',\n",
+    "       'ece_flag_number', 'cwr_flag_number', 'ack_count',\n",
+    "       'syn_count', 'fin_count', 'urg_count', 'rst_count', \n",
+    "    'HTTP', 'HTTPS', 'DNS', 'Telnet', 'SMTP', 'SSH', 'IRC', 'TCP',\n",
+    "       'UDP', 'DHCP', 'ARP', 'ICMP', 'IPv', 'LLC', 'Tot sum', 'Min',\n",
+    "       'Max', 'AVG', 'Std', 'Tot size', 'IAT', 'Number', 'Magnitue',\n",
+    "       'Radius', 'Covariance', 'Variance', 'Weight', \n",
+    "]\n",
+    "y_column = 'label'"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "249673a6-4826-4b80-b9aa-dfa4c3d549c4",
+   "metadata": {},
+   "source": [
+    "### Scaling"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 5,
+   "id": "cba40f31",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "from sklearn.preprocessing import MinMaxScaler, StandardScaler\n",
+    "scaler = StandardScaler()"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 6,
+   "id": "3682559f-9eb3-4d35-b1b2-d7d501ab85bc",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "for train_set in tqdm(training_sets):\n",
+    "    scaler.fit(pd.read_csv(DATASET_DIRECTORY + train_set)[X_columns])"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "60abc3f0-e32d-40be-abc5-fd5972cf9856",
+   "metadata": {},
+   "source": [
+    "### Classification: 34 (33+1) classes"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 7,
+   "id": "d208cf46-8ba9-480f-ab99-d4ee81c083b4",
+   "metadata": {
+    "tags": []
+   },
+   "outputs": [],
+   "source": [
+    "ML_models = [\n",
+    "        LogisticRegression(n_jobs=-1),\n",
+    "]\n",
+    "\n",
+    "ML_neams = [\n",
+    "        \"LogisticRegression\",\n",
+    "]\n",
+    "\n",
+    "for train_set in tqdm(training_sets):\n",
+    "    d = pd.read_csv(DATASET_DIRECTORY + train_set)\n",
+    "    d[X_columns] = scaler.transform(d[X_columns])\n",
+    "    for model in (ML_models):\n",
+    "        model.fit(d[X_columns], d[y_column])\n",
+    "    del d"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 8,
+   "id": "6116132e-02f0-4bac-aefb-2ba0bee924ab",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "y_test = []\n",
+    "preds = {i:[] for i in range(len(ML_models))}\n",
+    "for test_set in tqdm(test_sets):\n",
+    "    d_test = pd.read_csv(DATASET_DIRECTORY + test_set)\n",
+    "    d_test[X_columns] = scaler.transform(d_test[X_columns])\n",
+    "    \n",
+    "    y_test += list(d_test[y_column].values)\n",
+    "    \n",
+    "    for i in range(len(ML_models)):\n",
+    "        model = ML_models[i]\n",
+    "        y_pred = list(model.predict(d_test[X_columns]))\n",
+    "        preds[i] = preds[i] + y_pred\n",
+    "        "
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 9,
+   "id": "375dcbfb-2b20-4b37-8fbb-c9d68a6ac541",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "from sklearn.metrics import accuracy_score, recall_score, precision_score, f1_score\n",
+    "for k,v in preds.items():\n",
+    "    y_pred = v\n",
+    "    print(f\"##### {ML_neams[k]} (34 classes) #####\")\n",
+    "    print('accuracy_score: ', accuracy_score(y_pred, y_test))\n",
+    "    print('recall_score: ', recall_score(y_pred, y_test, average='macro'))\n",
+    "    print('precision_score: ', precision_score(y_pred, y_test, average='macro'))\n",
+    "    print('f1_score: ', f1_score(y_pred, y_test, average='macro'))\n",
+    "    print()\n",
+    "    print()\n",
+    "    print()"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "3958c6fa-6d05-48fb-a046-55e5843e4711",
+   "metadata": {},
+   "source": [
+    "# Classification: 8 (7+1) classes"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 10,
+   "id": "9208c899-8b57-4a3a-a2e7-94b057123536",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "dict_7classes = {}\n",
+    "dict_7classes['DDoS-RSTFINFlood'] = 'DDoS'\n",
+    "dict_7classes['DDoS-PSHACK_Flood'] = 'DDoS'\n",
+    "dict_7classes['DDoS-SYN_Flood'] = 'DDoS'\n",
+    "dict_7classes['DDoS-UDP_Flood'] = 'DDoS'\n",
+    "dict_7classes['DDoS-TCP_Flood'] = 'DDoS'\n",
+    "dict_7classes['DDoS-ICMP_Flood'] = 'DDoS'\n",
+    "dict_7classes['DDoS-SynonymousIP_Flood'] = 'DDoS'\n",
+    "dict_7classes['DDoS-ACK_Fragmentation'] = 'DDoS'\n",
+    "dict_7classes['DDoS-UDP_Fragmentation'] = 'DDoS'\n",
+    "dict_7classes['DDoS-ICMP_Fragmentation'] = 'DDoS'\n",
+    "dict_7classes['DDoS-SlowLoris'] = 'DDoS'\n",
+    "dict_7classes['DDoS-HTTP_Flood'] = 'DDoS'\n",
+    "\n",
+    "dict_7classes['DoS-UDP_Flood'] = 'DoS'\n",
+    "dict_7classes['DoS-SYN_Flood'] = 'DoS'\n",
+    "dict_7classes['DoS-TCP_Flood'] = 'DoS'\n",
+    "dict_7classes['DoS-HTTP_Flood'] = 'DoS'\n",
+    "\n",
+    "\n",
+    "dict_7classes['Mirai-greeth_flood'] = 'Mirai'\n",
+    "dict_7classes['Mirai-greip_flood'] = 'Mirai'\n",
+    "dict_7classes['Mirai-udpplain'] = 'Mirai'\n",
+    "\n",
+    "dict_7classes['Recon-PingSweep'] = 'Recon'\n",
+    "dict_7classes['Recon-OSScan'] = 'Recon'\n",
+    "dict_7classes['Recon-PortScan'] = 'Recon'\n",
+    "dict_7classes['VulnerabilityScan'] = 'Recon'\n",
+    "dict_7classes['Recon-HostDiscovery'] = 'Recon'\n",
+    "\n",
+    "dict_7classes['DNS_Spoofing'] = 'Spoofing'\n",
+    "dict_7classes['MITM-ArpSpoofing'] = 'Spoofing'\n",
+    "\n",
+    "dict_7classes['BenignTraffic'] = 'Benign'\n",
+    "\n",
+    "dict_7classes['BrowserHijacking'] = 'Web'\n",
+    "dict_7classes['Backdoor_Malware'] = 'Web'\n",
+    "dict_7classes['XSS'] = 'Web'\n",
+    "dict_7classes['Uploading_Attack'] = 'Web'\n",
+    "dict_7classes['SqlInjection'] = 'Web'\n",
+    "dict_7classes['CommandInjection'] = 'Web'\n",
+    "\n",
+    "\n",
+    "dict_7classes['DictionaryBruteForce'] = 'BruteForce'"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 11,
+   "id": "4c1f697f-88d8-4ac4-8bc6-f1a8ac3794d5",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "from sklearn.ensemble import RandomForestClassifier\n",
+    "from sklearn.linear_model import LogisticRegression\n",
+    "\n",
+    "\n",
+    "ML_models = [\n",
+    "        LogisticRegression(n_jobs=-1),\n",
+    "]\n",
+    "\n",
+    "ML_neams = [\n",
+    "        \"LogisticRegression\",\n",
+    "]\n",
+    "\n",
+    "\n",
+    "for train_set in tqdm(training_sets):\n",
+    "    d = pd.read_csv(DATASET_DIRECTORY + train_set)\n",
+    "    d[X_columns] = scaler.transform(d[X_columns])\n",
+    "    new_y = [dict_7classes[k] for k in d[y_column]]\n",
+    "    d[y_column] = new_y\n",
+    "    \n",
+    "    for model in (ML_models):\n",
+    "        model.fit(d[X_columns], d[y_column])\n",
+    "    del d"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 12,
+   "id": "6b69c509-7666-45bd-9e11-52ecec0df8a8",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "y_test = []\n",
+    "preds = {i:[] for i in range(len(ML_models))}\n",
+    "for test_set in tqdm(test_sets):\n",
+    "    d_test = pd.read_csv(DATASET_DIRECTORY + test_set)\n",
+    "    d_test[X_columns] = scaler.transform(d_test[X_columns])\n",
+    "    new_y = [dict_7classes[k] for k in d_test[y_column]]\n",
+    "    d_test[y_column] = new_y\n",
+    "    \n",
+    "    y_test += list(d_test[y_column].values)\n",
+    "    \n",
+    "    for i in range(len(ML_models)):\n",
+    "        model = ML_models[i]\n",
+    "        y_pred = list(model.predict(d_test[X_columns]))\n",
+    "        preds[i] = preds[i] + y_pred\n",
+    "        "
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 13,
+   "id": "3e0a9702-63f5-4898-a8b0-2bf950fe881d",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "from sklearn.metrics import accuracy_score, recall_score, precision_score, f1_score\n",
+    "for k,v in preds.items():\n",
+    "    y_pred = v\n",
+    "    print(f\"##### {ML_neams[k]} (8 classes) #####\")\n",
+    "    print('accuracy_score = ', accuracy_score(y_pred, y_test))\n",
+    "    print('recall_score = ', recall_score(y_pred, y_test, average='macro'))\n",
+    "    print('precision_score = ', precision_score(y_pred, y_test, average='macro'))\n",
+    "    print('f1_score = ', f1_score(y_pred, y_test, average='macro'))\n",
+    "    print()\n",
+    "    print()\n",
+    "    print()"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "a6ecac59-fc02-4198-9910-daf890da7a0a",
+   "metadata": {},
+   "source": [
+    "# Classification: 2 (1+1) Classes"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 14,
+   "id": "90ee4a99-d160-43bc-b2a0-06fa3f49e222",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "dict_2classes = {}\n",
+    "dict_2classes['DDoS-RSTFINFlood'] = 'Attack'\n",
+    "dict_2classes['DDoS-PSHACK_Flood'] = 'Attack'\n",
+    "dict_2classes['DDoS-SYN_Flood'] = 'Attack'\n",
+    "dict_2classes['DDoS-UDP_Flood'] = 'Attack'\n",
+    "dict_2classes['DDoS-TCP_Flood'] = 'Attack'\n",
+    "dict_2classes['DDoS-ICMP_Flood'] = 'Attack'\n",
+    "dict_2classes['DDoS-SynonymousIP_Flood'] = 'Attack'\n",
+    "dict_2classes['DDoS-ACK_Fragmentation'] = 'Attack'\n",
+    "dict_2classes['DDoS-UDP_Fragmentation'] = 'Attack'\n",
+    "dict_2classes['DDoS-ICMP_Fragmentation'] = 'Attack'\n",
+    "dict_2classes['DDoS-SlowLoris'] = 'Attack'\n",
+    "dict_2classes['DDoS-HTTP_Flood'] = 'Attack'\n",
+    "\n",
+    "dict_2classes['DoS-UDP_Flood'] = 'Attack'\n",
+    "dict_2classes['DoS-SYN_Flood'] = 'Attack'\n",
+    "dict_2classes['DoS-TCP_Flood'] = 'Attack'\n",
+    "dict_2classes['DoS-HTTP_Flood'] = 'Attack'\n",
+    "\n",
+    "\n",
+    "dict_2classes['Mirai-greeth_flood'] = 'Attack'\n",
+    "dict_2classes['Mirai-greip_flood'] = 'Attack'\n",
+    "dict_2classes['Mirai-udpplain'] = 'Attack'\n",
+    "\n",
+    "dict_2classes['Recon-PingSweep'] = 'Attack'\n",
+    "dict_2classes['Recon-OSScan'] = 'Attack'\n",
+    "dict_2classes['Recon-PortScan'] = 'Attack'\n",
+    "dict_2classes['VulnerabilityScan'] = 'Attack'\n",
+    "dict_2classes['Recon-HostDiscovery'] = 'Attack'\n",
+    "\n",
+    "dict_2classes['DNS_Spoofing'] = 'Attack'\n",
+    "dict_2classes['MITM-ArpSpoofing'] = 'Attack'\n",
+    "\n",
+    "dict_2classes['BenignTraffic'] = 'Benign'\n",
+    "\n",
+    "dict_2classes['BrowserHijacking'] = 'Attack'\n",
+    "dict_2classes['Backdoor_Malware'] = 'Attack'\n",
+    "dict_2classes['XSS'] = 'Attack'\n",
+    "dict_2classes['Uploading_Attack'] = 'Attack'\n",
+    "dict_2classes['SqlInjection'] = 'Attack'\n",
+    "dict_2classes['CommandInjection'] = 'Attack'\n",
+    "\n",
+    "dict_2classes['DictionaryBruteForce'] = 'Attack'"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 15,
+   "id": "506eae35-a310-4a34-8bcf-c99282ed3225",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "from sklearn.ensemble import RandomForestClassifier\n",
+    "from sklearn.linear_model import LogisticRegression\n",
+    "\n",
+    "\n",
+    "ML_models = [\n",
+    "        LogisticRegression(n_jobs=-1),\n",
+    "]\n",
+    "\n",
+    "ML_neams = [\n",
+    "        \"LogisticRegression\",\n",
+    "]\n",
+    "\n",
+    "\n",
+    "for train_set in tqdm(training_sets):\n",
+    "    d = pd.read_csv(DATASET_DIRECTORY + train_set)\n",
+    "    d[X_columns] = scaler.transform(d[X_columns])\n",
+    "    new_y = [dict_2classes[k] for k in d[y_column]]\n",
+    "    d[y_column] = new_y\n",
+    "    \n",
+    "    for model in (ML_models):\n",
+    "        model.fit(d[X_columns], d[y_column])\n",
+    "    del d"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 16,
+   "id": "b07aa379-ec7e-4651-ab5a-6845ae249132",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "y_test = []\n",
+    "preds = {i:[] for i in range(len(ML_models))}\n",
+    "for test_set in tqdm(test_sets):\n",
+    "    d_test = pd.read_csv(DATASET_DIRECTORY + test_set)\n",
+    "    d_test[X_columns] = scaler.transform(d_test[X_columns])\n",
+    "    new_y = [dict_2classes[k] for k in d_test[y_column]]\n",
+    "    d_test[y_column] = new_y\n",
+    "    \n",
+    "    y_test += list(d_test[y_column].values)\n",
+    "    \n",
+    "    for i in range(len(ML_models)):\n",
+    "        model = ML_models[i]\n",
+    "        y_pred = list(model.predict(d_test[X_columns]))\n",
+    "        preds[i] = preds[i] + y_pred\n",
+    "        "
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 17,
+   "id": "caabf4fd-097d-4db2-847a-0dcd87144d6f",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "from sklearn.metrics import accuracy_score, recall_score, precision_score, f1_score\n",
+    "for k,v in preds.items():\n",
+    "    y_pred = v\n",
+    "    print(f\"##### {ML_neams[k]} (2 classes) #####\")\n",
+    "    print('accuracy_score: ', accuracy_score(y_pred, y_test))\n",
+    "    print('recall_score: ', recall_score(y_pred, y_test, average='macro'))\n",
+    "    print('precision_score: ', precision_score(y_pred, y_test, average='macro'))\n",
+    "    print('f1_score: ', f1_score(y_pred, y_test, average='macro'))\n",
+    "    print()\n",
+    "    print()\n",
+    "    print()"
+   ]
+  }
+ ],
+ "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": "3.9.12"
+  }
+ },
+ "nbformat": 4,
+ "nbformat_minor": 5
+}