Joint Security–Performance Intelligence through Dual-Objective Learning for Autonomous Next-Generation Networks

Abstract

The rapid evolution of 6G communication networks has introduced unprecedented challenges in ensuring secure, adaptive, and high-performance network operations. Traditional intrusion detection and network monitoring techniques, which primarily rely on static rules and manual analysis, are increasingly ineffective against sophisticated cyberattacks and dynamic traffic patterns. An intelligent Machine Learning (ML)-based framework is developed to perform dual-purpose analysis by simultaneously classifying network attacks and predicting network throughput using Classification and Regression Tree (CART)- based learning. The framework integrates Support Vector Machine (SVM), k-Nearest Neighbors (KNN), and a Tree-based Adaptive Optimization (TAO) ensemble model to improve detection capability and predictive accuracy. Data preprocessing techniques, including Label Encoding, feature standardization, and Synthetic Minority Over-sampling Technique (SMOTE), are employed to enhance data quality, reduce class imbalance, and improve model generalization. Experimental evaluation demonstrates that the TAO ensemble model outperforms conventional ML algorithms by achieving higher attack classification accuracy and more reliable throughput prediction across diverse network conditions. A Flask-based web application enables real-time traffic analysis, throughput prediction, and interactive visualization, facilitating practical deployment in intelligent network management systems. The proposed framework provides a scalable, adaptive, and computationally efficient solution for strengthening cybersecurity while optimizing communication performance, making it suitable for secure and intelligent resource management in future 6G network environments.