This project combines AI-based modeling and scripted tools to detect intrusions in wireless networks using packet capture (.pcap) data. It focuses on the AWID dataset, applying machine learning techniques for training models and using them to detect malicious network activity.
A comprehensive exploratory data analysis (EDA) and model training pipeline is provided using the AWID dataset. This process includes data cleaning, feature selection, and training a Random Forest classifier.
👉 AWID Dataset EDA & Model Training Notebook
The notebook includes:
- Dataset exploration and visualization
- Preprocessing and feature engineering
- Model training and evaluation
- Exporting the trained model (
rf_model.pkl) for use in detection scripts
These scripts use the trained model to analyze .pcap files and detect intrusions either offline or in real time.
Supports parallel and batch processing for large .pcap files.
# For parallel processing (recommended for large files)
python ids_multi_process.py --pcap wpa2-cracking-phone-access-point.pcap --model rf_model.pkl --parallel
# For batch processing
python ids_multi_process.py --pcap wpa2-cracking-phone-access-point.pcap --model rf_model.pkl --batch --batch-size 500
# To limit the number of packets processed (with parallel processing)
python ids_multi_process.py --pcap wpa2-cracking-phone-access-point.pcap --model rf_model.pkl --parallel --limit 10000A basic script for quick and simple intrusion detection.
python simple_ids.pyPerforms real-time intrusion detection on wireless network traffic.
Install all dependencies before running the scripts:
pip install -r requirements.txt