MACHINE LEARNING MODEL TO PREDICT SNR IN 6G IOT NETWORKS

Abstract

The emergence of sixth-generation (6G) wireless networks is expected to reshape communication by delivering ultra-reliable, low-latency, and high-capacity connections for large-scale Internet of Everything (IoE) scenarios. Yet, sustaining a strong signal-to-noise ratio (SNR) in terahertz (THz) and visible light communication (VLC) bands remains difficult due to severe propagation loss and environmental variability. This work applies machine learning classification methods—including Logistic Regression, Random Forest, and Multilayer Perceptron (MLP)—to the “6G IoT Intelligent Management Dataset” to examine beamforming optimization across varied conditions. Using feature engineering, SHAP-based interpretability, and empirical testing, the results reveal that beamforming gain and SNR improvement are key indicators of communication quality. The MLP delivers the best performance with an F1 score of 0.92, underscoring the promise of deep learning in adaptive transmission control. These outcomes highlight the role of context-aware, AI-enabled methods in addressing 6G reliability issues and supporting sustainable IoT connectivity.