Predicting QoE of Video Streaming with Network-Context Features
Han Hu, Zhimin Xu, Xinggong Zhang, Zongming Guo · 2019
It is always the top priority for network service providers to provide better Quality of Experience (QoE). Huge efforts have been devoted to Adaptive Bit Rate (ABR) streaming, but it is only a passive QoE amelioration method by adapting to bandwidth variations. In this paper, we aim to provide a QoE prediction tool for service providers to improve QoE proactively. We have investigated the impacts of various network conditions on QoE through a large-scale measurement study, and proposed a QoE model from network-context features. The correlation analysis helps to figure out how the network-context features influence QoE. Furthermore, we designed a neural-network learning algorithm to predict QoE with the features of access link, location, CDN and network operators. The prediction algorithm achieves up to 85% accuracy over a large dataset of 2 billion viewing sessions from a commercial video website.