Online playtime prediction for cognitive video streaming

In this paper, we consider the problem of cognitive video streaming in video on demand (VoD) services. The focus lies on quantities that are indicative of the quality of experience (QoE) of the subscriber, such as playtime ratio, probability of return, probability of replay and startup time. Especially, in this paper, we develop and evaluate a playtime prediction tool. For this purpose, the applicability of different machine learning algorithms such as k-nearest neighbor, neural network regression, and survival models is investigated; then, we develop an approach to identify the most relevant factors that contributed to the prediction. The proposed approaches are tested by means of a data set provided by Comcast.