User Equipment Grouping in 5G TDD System using Machine Learning

Korkut Arslantürk ()

Abstract

The project's motivation is to show that traditional supervised and unsupervised machine learning methods can successfully be applied to UE channel measurements in a real-world commercial system to perform UE classification. This way we prove that UE classification is feasible also on real-life datasets. Results provided in this work may be useful input to the network when making UE grouping decisions as well as assisting input to the scheduling mechanisms. Moreover, the effect of UE classification on the best beam selection will be explored.