Predicting Unstable Handovers Using Machine Learning in LTE/5G Networks
Publication Date
1-1-2026
Document Type
Conference Proceeding
Publication Title
Proceedings of 2026 IEEE International Conference on Signals and Systems Icsigsys 2026
DOI
10.1109/ICSigSys71160.2026.11613879
First Page
26
Last Page
31
Abstract
Frequent and unnecessary handovers are a major challenge in dense cellular networks because they increase signaling overhead and reduce the benefit of mobility management. This project studies unstable handover prediction using a machine learning pipeline built on LTE simulation and mobility traces. The main target is defined as unstable_15 = reversal_15 OR short_dwell_15. We evaluate Logistic Regression, Explainable Boosting Machine (EBM), LightGBM, and XGBoost by grouping train-test splits by UE and performing offline gating analysis. Results show that boosted-tree models perform better than the linear baseline on ranking metrics, and LightGBM has the strongest balanced threshold tradeoff for offline gating. The study shows that pre-handover measurements can be used to identify a notable amount of unstable handover events. It also highlights the difficulty of the overall task and how sensitive gating performance is to model choice and threshold selection.
Keywords
5G/LTE, offline gating, sensitivity analysis, supervised machine learning, Unstable handover
Department
Computer Science
Recommended Citation
Niyati Aggarwal, Pooja Shyamsundar, and Navrati Saxena. "Predicting Unstable Handovers Using Machine Learning in LTE/5G Networks" Proceedings of 2026 IEEE International Conference on Signals and Systems Icsigsys 2026 (2026): 26-31. https://doi.org/10.1109/ICSigSys71160.2026.11613879
