Machine learning-based QoE prediction for video streaming over LTE network

This paper presents a machine-learning software solution that performs a multi-dimensional prediction of QoE (Quality of Experience) based on network-related SIFs (System Influence Factors) as input data. The proposed solution is verified through experimental study based on video streaming emulation over LTE (Long Term Evolution) which allows the measurement of network-related SIF (i.e., delay, jitter, loss), and subjective assessment of MOS (Mean Opinion Score). Obtained results show good performance of proposed MOS predictor in terms of mean prediction error and thereby can serve as an encouragement to implement such solution in all-IP (Internet Protocol) real environment.