akanksham23/ai-network-intrusion-detection

🛡️ AI-Based Network Intrusion Detection System (AI-NIDS)

An AI-powered Network Intrusion Detection System that combines Machine Learning (Random Forest) with Generative AI (Groq LLM) to detect, analyze, and explain network attacks, focusing on DDoS traffic from the CIC-IDS2017 dataset.

This project is designed for cybersecurity students to understand how traditional ML and modern LLMs can work together in Security Operations Centers (SOC).

🚀 How to Use

1. 1.Enter Groq API Key (Optional) Add your Groq API key in the sidebar to enable AI-based explanations.

1. 1.Train the AI Model Click “Train AI Model”. The app automatically loads the CIC-IDS2017 DDoS dataset.

1. 1.Simulate Network Traffic Click “Simulate Random Packet” to analyze a real network flow.

1. 1.Analyze Results
2. 2.Detects whether traffic is BENIGN or DDoS
3. 3.Displays confidence score
4. 4.Uses Groq AI to explain why the packet was classified as such

🧠 Key Features

• —✅ Random Forest–based intrusion detection
• —📊 Feature importance visualization
• —📈 Accuracy & performance metrics
• —🎯 Live packet simulation
• —🤖 LLM-based attack explanation (Groq)
• —🛡️ SOC-style security reasoning
• —🎓 Student-friendly explanations

📂 Project Structure

<pre> AI-Based Network Intrusion Detection System ├── app.py ├── requirements.txt ├── Friday-WorkingHours-Afternoon-DDos.pcap_ISCX.csv └── README.md </pre>

📄 File Details

• —app.py – Streamlit application code
• —requirements.txt – Python dependencies
• —CSV Dataset – CIC-IDS2017 DDoS traffic subset

📊 Dataset Used

• —CIC-IDS2017
• —Realistic labeled network traffic
• —Industry-standard dataset for IDS research

🎓 Learning Outcomes

• —Understand ML-based intrusion detection
• —Interpret network traffic features
• —Apply explainable AI in cybersecurity
• —Bridge ML detection with SOC decision-making
• —Gain hands-on experience with Streamlit dashboards

🧪 Future Enhancements

• —Real-time PCAP capture
• —Anomaly-based zero-day detection
• —Deep Learning (LSTM / Autoencoders)
• —MITRE ATT&CK mapping
• —PDF attack reports

👩‍💻 Author

Akanksha Vikas Mane BE IT | AI-Driven Security Research

⚠️ Disclaimer

This project is for educational and research purposes only. Not intended for direct production deployment.