Intrusion-Detection-System

This project implements a Machine Learning-based Intrusion Detection System (IDS) using the UNSW-NB15 dataset. It classifies network traffic as normal or malicious using a trained model, and serves predictions via a Flask API.

📁 Project Structure

ids-flask ├── app.py # Flask API to serve the ML model ├── train_test.ipynb # Notebook to preprocess data and train the model ├── unsw_ids_model.pkl # Trained model file (XGBoost) ├── requirements.txt # Python dependencies ├── post_request.py # Script to test POST request to Flask API ├── get_request.py # Script to test GET request to Flask API ├── get_post_response.jpg # Screenshot showing GET/POST 200 OK responses

⚙️ Requirements

• Install all dependencies using:

pip install -r requirements.txt # Check the requirements.txt f

📊 Model Training

The model is trained using the train_test.ipynb notebook. Key steps:

• Dataset preprocessing

• Train/test split

• Model training using XGBoostClassifier

• Model is saved as unsw_ids_model.pkl using pickle

🚀 Running the Flask API

-Start the API using:

python app.py

If successful, it will be hosted at: http://127.0.0.1:5000/

Available Routes:

GET / → Returns a simple message confirming the server is running.

POST /predict → Takes input features and returns prediction (0 for normal, 1 for intrusion).

📡 Testing the API

• For GET Request, use: get_request.py

• For POST Request, use: post_request.py

🖼️ See get_post_response.jpg for successful 200 OK responses , the image is attached.

✅ Output

• If the model predicts an intrusion, the response will be:

{ "prediction": 1 }

• If normal traffic:

{ "prediction": 0 }

If there is a mismatch in features (like missing attack_cat), an error will be returned — but this is okay for internship submission as long as the server runs and accepts requests.

📌 Notes

This project is not production-grade.

Designed as an academic submission for an AI Internship Project.

Data preprocessing and model training were done on Google Colab.