A QoE Estimation Model Considering Video Popularity for Video Streaming Services

With the increasing demand of keeping competitive force for video streaming service provider, the user-oriented QoE evaluation is significant to be optimized. In previous work, those user-centric QoE evaluation methods depend on user perception collection which are generally based on subjective feedbacks which causes the impracticality of evaluation model. In this paper, a QoE estimated model considering video popularity is proposed. The model substitutes the user preference by the mapping relationship between video popularity and user preference so that the service provider can estimate user perceived QoE with objective video information. The mapping relationship between video popularity and user preference is explored by analysis of uncertainty of user prefenrece for different popularity degree. Then, a crowdsourcing-based experiment is conducted to collect data for QoE model regression and verification. With data analysis, a QoE model based on video popularity is developed and the correlation coefficient shows the well estimation on QoE.