🛡️ One²: An Intrusion Detection System for Both Internal and External Vehicular Network from Weak Labeled Data

A PyTorch implementation of One² for intrusion detection across multiple IVN and EVN datasets.

📁 Project Structure

├── main.py                 # Main entry point
├── data_processor.py       # Data preprocessing and graph construction  
├── models.py              # HGT model definition
├── trainer.py             # Training and evaluation
├── utils.py               # Utility functions
├── dataset/               # Dataset files
├── model/                 # Saved checkpoints
├── output/                # Output files

🛠️ Environment Setup

Create Conda Environment

# Create a new conda environment
conda create -n hgt-ids python=3.8 -y

# Activate the environment
conda activate hgt-ids

Install from requirements.txt

pip install -r requirements.txt

Sample requirements.txt:

absl-py==2.1.0
aiohttp==3.9.3
aiosignal==1.3.1
asttokens==3.0.0
async-timeout==4.0.3
attrs==23.2.0
backcall==0.2.0

🚀 Run Commands

Basic Training

# Train model on CIC-IDS2017 dataset (binary classification)
python main.py --dataset CIC-IDS2017 --binary True --epochs 20

Advanced Configuration

# Custom hyperparameters
python main.py \
    --dataset TON_IoT \
    --file df.csv \
    --binary True \
    --batchsize 128 \
    --epochs 100 \
    --learning_rate 0.01 \
    --hidden_channels 64 \
    --num_heads 2 \
    --num_layers 1 \
    --cuda cuda:0

# Evaluation only (no training)
python main.py --dataset CIC-IDS2017 --train_if False --roc True

Supported Datasets

Dataset	Command
CIC-IDS2017	--dataset CIC-IDS2017
TON_IoT	--dataset TON_IoT
CAR-HACKING	--dataset CAR-HACKING
CAN-intrusion	--dataset CAN-intrusion
CIC-UNSW-NB15	--dataset CIC-UNSW-NB15
CICIoV2024	--dataset CICIoV2024

Command Line Arguments

Argument	Type	Default	Description
--dataset	str	CIC-IDS2017	Experimental dataset
--file	str	df.csv	Dataset file name
--binary	bool	True	Binary or multiclass classification
--batchsize	int	128	Batch size for training
--epochs	int	1	Number of training epochs
--learning_rate	float	0.01	Learning rate
--hidden_channels	int	64	Hidden layer dimensions
--num_heads	int	2	Number of attention heads
--num_layers	int	1	Number of HGT layers
--cuda	str	cuda:0	CUDA device
--train_if	bool	True	Whether to train the model
--roc	bool	True	Generate ROC curves

📊 Results and Outputs

• Model checkpoints: Saved in ./model/ directory
• Training metrics: CSV files in ./output/ directory
• ROC curves: Data saved in ./roc/ directory
• TensorBoard logs: Available in ./logs/ directory

🤝 Contributing

Feel free to submit issues and pull requests to improve this implementation!

📄 License

This project is open source. Please cite appropriately if used in research.