Enhancing 5G network performance through advanced network slicing: A Comprehensive evaluation using machine learning techniques
Abstract
This study explores the potential of network slicing to enhance 5G telecom network performance. By determining end-to-end network resource allocation based on service needs, programmable network slicing is proposed. Utilizing a comprehensive dataset of network architecture and service requirements, the study employs decision trees, neural networks, and clustering techniques to evaluate and optimize network slicing configurations. The analysis aims to uncover patterns and correlations in the dataset to enhance the network slicing process. Key performance indicators encompass metrics like bandwidth, coverage, security, latency, and reliability. The findings demonstrate that network slicing greatly enhances the velocity and efficiency of 5G apps and services. The logistic regression model used for threat detection had high accuracy (77.8%), precision (82.9%), recall (68.8%), and F1-score (65.4%). The Singular Value Decomposition (SVD) model had 76.6% accuracy, 56.5% precision, 66.7% recall, and 60.7% F1-score. These high metrics suggest effective real-time network control. The systematic approach entails a series of processes, including data preparation, feature engineering, model training, and evaluation. This article highlights the significance of strong security standards, flexible and interoperable network slicing technology, and innovative optimization methods. Network providers and mobile operators can greatly benefit from the strategic adoption of network slicing, which allows them to maximize the advantages of 5G networks. The findings derived from this research offer opportunities for further exploration of advanced network slicing technologies and their broader use in the evolving telecommunications industry.