Building a Network Intrusion Detection System

Overview

#

A machine learning based Network Intrusion Detection System built to classify network traffic as normal or malicious.

Tech Stack

#

• Classifiers: Random Forest, SVM, KNN via scikit-learn
• Clustering: K-Means with PCA for dimensionality reduction
• Feature Selection: Genetic Algorithm
• Dataset: NSL-KDD

How It Works

#

1. Raw network traffic captured and preprocessed
2. Features extracted and reduced via PCA
3. Genetic Algorithm selects optimal feature subset
4. Classifier trained and evaluated

Results

#

Achieved high classification accuracy with reduced feature set via genetic algorithm optimization.