Optimizing Network Routing for Security and Efficiency through Machine Learning Techniques

Abstract

This study presents contemporary communication systems, it is essential to regulate network traffic in a manner that is both efficient and secure. Many routing algorithms exhibit issues such as insufficient accuracy, prolonged processing times, inability to manage high traffic volumes, lack of security, and inadequate real-world testing. This study proposes an enhanced route selection algorithm that employs machine learning to optimise routing efficiency, enhance detection accuracy, and elevate overall network performance. constructed a customised dataset by emulating a network comprising both legitimate and malicious traffic. Also trained and evaluated four machine learning models: Decision Tree, Logistic Regression, Random Forest, and Support Vector Machine (SVM). Employed significant performance metrics to do this. The most efficient model was Random Forest, with the highest accuracy (96.86%), detection efficiency (98.64%), and a significantly reduced stolen packet rate of 1.00%. It demonstrated superior network performance with a packet delivery rate of 72.40%, reduced average hops, and enhanced path utilisation. The Random Forest-based method effectively identified assaults by accurately detecting malicious behaviour with little false negatives. The results indicate that machine learning-based routing could revolutionise the field, with Random Forest providing the optimal equilibrium among accuracy, security, and computational efficiency. The proposed design significantly enhances traffic management, facilitates scalability, and strengthens security. This addresses significant research deficiencies and paves the way for intelligent, practical network traffic control systems.