Research of users' viewing habits based on clustering method

摘要

With the development of value-added services in cable digital TV, user behavior analysis has important practical significance in improving stickiness and implementation of personalized marketing. Based on real viewing data collected by software in set-top box, this paper adopts clustering method to analyze viewing habit in a particular region of Shanxi Province. The results show that the prime time in this region is from 7PM to 11PM, and the difference between overall user viewing on workday and weekend is not significant. Users' favorite viewing hour is from 9PM to 10PM and users who watch TV longer and more frequently are more accustomed to watching TV in prime time.