An Algorithm of Users Access Patterns Mining Based on Video Recommendation

摘要

Due to the substantial growth of IPTV video users, the weak characteristics of terminal "Set-Top-Box + TV" lead to a great difficulty for users to search required videos, and effective recommendation has an important influence on improving VOD quality. In this paper, the Forecast algorithm is proposed, which is based on the sequential pattern method, consider timeliness features of the videos, convert user watch history into relative access value, then do personalized recommendation. It has strong response speed, can basically meet the needs of real-time video recommendation on IPTV. The algorithm is easy to understand, has good recommend result.