|
|
--- |
|
|
title: Network Intrusion Detection |
|
|
emoji: 🚨 |
|
|
colorFrom: red |
|
|
colorTo: blue |
|
|
sdk: gradio |
|
|
sdk_version: 3.50.0 |
|
|
app_file: app.py |
|
|
pinned: false |
|
|
--- |
|
|
|
|
|
# 🚨 Network Intrusion Detection System |
|
|
|
|
|
A real-time AI-powered system that detects malicious network traffic and classifies attack types using a CNN model trained on the KDD Cup 1999 dataset. |
|
|
|
|
|
## 🔍 Features |
|
|
|
|
|
- **Real-time detection** of 40+ attack types |
|
|
- **Multi-class classification** (not just binary) |
|
|
- **High accuracy** (~97% on test data) |
|
|
- **Attack type identification** with confidence scores |
|
|
|
|
|
## 🛡️ Attack Types Detected |
|
|
|
|
|
- **DoS Attacks**: neptune, smurf, teardrop, pod, back |
|
|
- **Probing Attacks**: portsweep, ipsweep, nmap, satan |
|
|
- **R2L Attacks**: guess_passwd, warezclient, imap, ftp_write |
|
|
- **U2R Attacks**: buffer_overflow, rootkit, loadmodule |
|
|
|
|
|
## 🚀 Usage |
|
|
|
|
|
1. Enter network traffic features |
|
|
2. Click "Submit" |
|
|
3. Get instant classification with confidence scores |
|
|
4. View top 3 predictions |
|
|
|
|
|
## 📊 Model Performance |
|
|
|
|
|
- Test Accuracy: ~97% |
|
|
- Multi-class F1 Score: ~96% |
|
|
- Training Data: KDD Cup 1999 dataset |
|
|
- Model: Custom CNN with 3 convolutional layers |
|
|
|
