Intelligent Network Traffic Anomaly Detection Using Machine Learning Algorithms

Abstract

With the rapid growth of internet communication and cloud-based services, network security has become a critical challenge due to the increasing number of cyber threats and malicious attacks. Traditional intrusion detection systems rely mainly on predefined signatures and rules, making them less effective against modern and evolving attacks. To address these limitations, this paper presents an intelligent network traffic anomaly detection system using machine learning algorithms. The proposed system utilizes the KDD Cup 1999 dataset for training and evaluation. Data preprocessing techniques such as feature encoding and feature selection are applied to improve prediction performance. Two machine learning algorithms, Random Forest and CATBoost Classifier, are implemented for anomaly detection. The trained models are integrated into a Flask-based web application that supports both CSV file upload analysis and real-time manual prediction. Experimental results demonstrate that the proposed system achieves high accuracy exceeding 99% while maintaining reliable prediction performance. The system also provides visualization features for improved analysis and interpretation of network traffic behaviour. Overall, the proposed approach offers a scalable, efficient, and user-friendly solution for intelligent intrusion detection and network security enhancement.