Effective Next-Items Recommendation via Personalized Sequential Pattern Mining

Abstract. Based on the intuition that frequent patterns can be used to predict the next few items that users would want to access, sequential pattern mining-based next-items recommendation algorithms have performed well in empirical stud-ies including online product recommendation. However, most current methods do not perform personalized sequential pattern mining, and this seriously limits their capability to recommend the best next-items to each specific target user. In this paper, we introduce a personalized sequential pattern mining-based recom-mendation framework. Using a novel Competence Score measure, the proposed framework effectively learns user-specific sequence importance knowledge, and exploits this additional knowledge for accurate personalized recommendation. Experimental results on real-world datasets demonstrate that the proposed frame-work effectively improves the efficiency for mining sequential patterns, increases the user-relevance of the identified frequent patterns, and most importantly, gener-ates significantly more accurate next-items recommendation for the target users.