Resampling Techniques for Enhanced Network Slice Classification in 5G Networks A SMOTE-Tomek Perspective

In the rapidly evolving landscape of 5G networks, the classification of network slices plays a pivotal role in optimizing resource allocation and providing quality of service to various connected devices. This study focuses on leveraging a 5G-enabled network slice dataset collected from diverse devices and employs a range of machine learning models to classify these network slices accurately. The challenge of class imbalance, inherent in many real-world datasets, is addressed through the application of the SMOTETomek method. Our research compares the performance of several machine learning models, including Random Forest, CatBoost, and Extra Trees, in classifying network slices. Our findings reveal that Decision Trees outshine other models, demonstrating superior performance in the classification of network slices in 5G networks. The Decision Tree model not only offers high accuracy but also ensures efficient resource allocation and management, making it an optimal choice for enhancing the capabilities of 5G networks and meeting the demands of various connected devices. This study serves as a valuable resource for network operators and researchers looking to maximize the potential of 5G technology. 

 Keywords: Decision Tree, Random Forest, Extra Tree, Catboost.