A Machine Learning Approach for Network Slice Selection

The surge in diverse and demanding applications in 5G networks requires advanced management techniques, such as network slicing, to ensure optimal service quality across varied requirements. This paper introduces a machine learning-based approach for optimizing network slice selection in 5G networks, focusing on the AdaBoostClassifier due to its robust predictive capabilities. Key performance indicators critical for slice differentiation such as packet loss rate and technology support are processed through extensive data preprocessing and feature engineering to improve the model's accuracy and efficiency. Furthermore, a data balancing method was implemented to mitigate the model's bias towards the dominant class. This approach significantly enhanced the model's overall performance, resulting in more accurate and reliable outcomes