Forecasting and Mitigating Interference in TDD Mobile Communication Networks Using Machine Learning: A Data-Driven Approach

In this paper, we investigate the application of artificial intelligence (AI) and machine learning (ML) methods to predict potential cross-link interference occurrences in Time Division Duplex (TDD) networks and identify optimal interference mitigation strategies. By combining network performance data with weather information, our objective is to establish predictive models capable of anticipating interference challenges and suggesting proactive solutions. Utilizing datasets obtained from a prominent Indian telecommunication provider, enriched with meteorological data, we train and evaluate the models. Through the implementation of a Random Forest (RF) algorithm, we achieve a remarkable prediction accuracy of $\mathbf{9 4. 0 7 \%}$ for aggressor nodes, $\mathbf{9 7. 6 \%}$ for victim nodes, and $\mathbf{9 3. 8 5 \%}$ for mitigation schemes.