Incorporating Pageview Weight into an Association-Rule-Based Web Recommendation System

Web recommendation systems based on web usage mining try to mine users' behavior patterns from web access logs, and recommend pages to the online user by matching the user's browsing behavior with the mined historical behavior patterns. Recommendation approaches proposed in previous works, however, do not distinguish the importance of different pageviews, and all the visited pages are treated equally whatever their usefulness to the user. We propose to use pageview duration to judge its usefulness to a user, and try to give more consideration to more useful pageviews, in order to better capture the user's information need and recommend pages with higher quality. In this paper we try to incorporate pageview weight into the Association Rule based model and develop a Weighted Association Rule model. Comparative experiment of the two shows a significant improvement in the recommendation effectiveness with the proposed Weighted Association Rule model.

• Type: Conference
• Year: 2006
• Where: AUSAI
• Authors: Liang Yan, Chunping Li
• Keywords: Artificial Intelligence, Association Rule Model, AUSAI 2006, Behavior Patterns, Weighted Association Rule