Comparison Of Machine Learning Algorithms For Predicting 5g Coverage Predictive Accuracy And The Identification Of Dominant Feature Parameters

Abstract

In 5G technology, the prediction of coverage areas plays a vital part in network optimization and reliable connectivity. In this paper, coverage area prediction is presented on an extensive comparative analysis involving many machine learning algorithms based on the RF Signal Data. The target column, Band Width, is used to determine prediction accuracy of various models through evaluation. Evaluation is performed using traditional methods like Logistic Regression, K-Nearest Neighbors (KNN), Naive Bayes, Random Forest, Support Vector Machine (SVM), XGBoost, LightGBM, AdaBoost, Bayesian Network Classifier, Multi-Layer Perceptron (MLP), Long Short-Term Memory (LSTM); against proposed advanced techniques such as Stacking and Voting Classifiers, and Convolutional Neural Networks (CNN). The aim is to find the feature parameters that strongly influence 5G coverage prediction. This research is intended to benchmark the performance and accuracy of these algorithms through developing a wide range of models. The comparative analysis provides the advantageous and disadvantageous factors for each methodology, thus giving valuable insights for researchers and network engineers. The conclusion drawn from this work is that ensemble methods, namely Stacking and Voting Classifiers, along with CNN, attained much higher prediction accuracies and robustness, and therefore, are viable solutions for improving 5G network planning and deployment.