A New Framework for Detecting Anomalies in Network Traffic Using Supervised Learning Techniques

Abstract

The new framework for detecting anomalies in network traffic is detailed model for novel network security, specifically as cyber threats resume to grow in intricacy. This work investigates Anomaly detection in network traffic is a critical component of modern network security, specifically as cyber threats continue to grow in intricacy. This framework investigates the efficiency of numerous supervised learning methods for classifying anomalies in network data, with a identifiable focus on their facility to handle challenges such as class imbalance and high dimensional feature spaces. The estimated methods include Isolation Forest, Naïve Bayes, Light GBM and Support Vector Machine (SVM) classifiers. Across complete investigation, both supervised and unsupervised techniques are analysed and contrasted using key performing metrics such as accuracy, precision and recall.The proposed anomaly detection framework follow four major stages: Preprocessing, Enhancement, feature Extraction and Selection. New results establish that the framework attains higher exactness and robust overall accuracy. Using the sample dataset, This study highlights the comparative strengths of supervised and unsupervisied models and provides a reliable framework for effective network anomaly detection.