Performance Evaluation of Machine Learning Regression Models for 5G Network Resource Allocation Optimization

This is the authors' abstract. We don't add key points for this paper.

A new era in wireless communication has begun with the arrival of fifth-generation [5G] technology, which promises unprecedented data rates, very low latency, and wide device connection. In order to maximize the effectiveness of 5G network resource allocation, this study suggests a method that makes use of ML. The study evaluates 5G network resource allocation through extensive ML model utilization. The 5G Resource Allocation Dataset features 8 attributes within 400 entries to assess signal quality and network performance optimization across different uses. An evaluation of four regression models, including K-Nearest Neighbors (KNN) and Random Forest Regressor (RFR), together with Gradient Boosting Regressor (GBR) and Extra Trees Regressor (ETR), was performed using Performance Metrics, which included R-squared, Mean Squared Error (MSE), and Root Mean Squared Error (RMSE) metrics. The KNN model provided the highest accuracy yield with the highest R-squared value of 0.9378 with a low MSE of about 0.0034 and RMSE of about 0.0583. More so, Random Forest Regressor model was also effective with an R-Squared of 0.8655. The analysis demonstrated that KNN and RFR achieved higher accuracy performance than GB and ETR.