Data-driven multi-criteria decision making for cell performance detection in 5G and B5G networks

Self-organizing communication networks in 5G and B5G automatically optimize and improve without human intervention, reducing OPEX and CAPEX. One of the critical features is self-healing, which predicts and resolves anomalies, optimizes performance, and adapts to changing cell density. This is managed through cell outage management (COM), which includes cell outage detection (COD) and cell outage compensation (COC). Timely prediction of cell state enhances network performance. This research uses machine learning techniques combined with multi-criteria decision making (MCDM) methods to rank and estimate defective cells. Using the TOPSIS method, an unsupervised dataset was transformed into a two-class supervised dataset. The results show that the TOPSIS-Adaboost combination outperforms other methods and improves the capabilities of cell outage detection in 5G and B5G networks.