A Comprehensive Evaluation of Time-Series Forecasting Methods for Proactive Handover

With the advancement of 5G and beyond, mobility management has become increasingly complex, driving the need for advanced proactive handover (HO) strategies. Traditionally, HO algorithms have been reactive, but recent advancements in AI/ML offer proactive solutions through time-series forecasting. In this paper, we present a comprehensive comparison of state-of-the-art time-series forecasting methods essential for proactive HO within the 3rd Generation Partnership Project simulation environment. Instead of focusing solely on overall prediction accuracy, we assess performance across various HO-related metrics to gain a thorough understanding. Additionally, we introduce a novel masking mechanism that prioritizes potential target cells and specific time points critical for HO decisions, enhancing the predictive power of time-series methods. This approach leads to improved HO performance and underscores the need for further exploration in this area.