DATA MINING-DRIVEN VERTICAL HANDOVER DECISION FRAMEWORK FOR 5G NETWORKS

Abstract

Effective and seamless vertical handover (VHO) is essential for sustained connectivity and high QoS in 5G heterogeneous networks. However, varying network behaviors and protocols complicate VHO decisions, leading to latency and service disruptions. This paper proposes a data mining-based VHO decision framework that leverages historical handover data using multivariate regression and ANOVA to identify key parameters such as signal strength, bandwidth, jitter, latency, packet loss, and coverage. Simulations in NetNeuman demonstrate improved network performance, reduced latency, and higher handover success rates compared to baseline methods. Real-time decisions based on historical trends enhance user experience and network reliability. Future integration with advanced machine learning could enable adaptive and predictive handovers for 6G networks, supporting ultra-reliable, low-latency communication.