Machine learning-based throughput prediction using communication quality in mobile networks

摘要

Throughput prediction contributes a lot to bitrate control technique, to improve user experience of video streaming. Existing method did not take various scenarios of user movement pattern into consideration in mobile networks. To bridge this gap, we put forward a throughput prediction model in this paper. The model first identifies the user movement pattern, then predicts throughput using communication quality data in the specific scenario with machine learning for mobile networks. Field experiment results indicate the method can predict throughput effectively in various scenarios.