Understanding User Behavior in Large Scale Internet Video Service

With the rapid development of Internet video service and the continuous expansion of user scale, to dig out the user behavior patterns accurately from the user data and obtain some valuable regularities of the user behavior can help the Internet video operators to improve service quality and enhance the user´s quality of experience (QoE), so that achieves the purpose of improving economic benefits. The goal of this paper is to provide a better understanding of the user behavior for video operators. To this end, we research the user behavior in the scene of large scale Internet video service based on the data set captured from one of the largest operators of China. By the way of statistical analysis and data mining of large scale user data, we conclude several regularities and statistical models of the user behavior. We also develop a deep study on the relevant factors that impact the user behavior, and focus on the relevance between the user behavior and video length, video popularity, buffer rate, error rate, which find out that the buffer rate have a significant impact on the user behavior, while the video popularity have a weak correlation with the user behavior.