Predicting 5G Coverage with Machine Learning: Dominant Features and Algorithmic Accuracy Comparison

For network performance enhancements, such prediction of coverage areas is critical to minimize the degradation and to ensure maximum connectivity for mobile users. In this scenario, the article thoroughly evaluates the Voting Classifier Analysis approach of predicting 5G coverage based on RF Signal Data. The target column Band Width is used for measuring the prediction accuracy. The core of this research is on modern ensemble techniques, specifically Voting Classifier, that arguably have more reliability in prediction as compared to existing ones. Given the advanced nature of ensemble analysis, it would be a challenge to identify very significant feature parameters for 5G coverage prediction. Benchmarking for such performance and accuracy will, therefore, be beneficial for network engineers and researchers alike. The results show that the Voting Classifier is capable of offering greater prediction accuracy and robustness as one of the best toward 5G planning and deployment optimizations.