Preference modeling for personalized retrieval based on browsing history analysis
Xiaoyu Zhang · IEEJ Transactions on Electrical and Electronic Engineering · 2013
Abstract Personalized retrieval aims at meeting the personalized information need of users, in which preference modeling is of great importance. The user's preference can be revealed via user specification and relevance feedback, both of which require extra effort from the user and are inevitably labor intensive. In this paper, we propose a novel preference modeling algorithm based on browsing history analysis for personalized retrieval. Based on the browsing log, we explore the user's interest in both fields and field values and accumulatively update the preference model. Given a query, the relevant retrieval results can spontaneously be ranked according to their corresponding preference score without any extra user interference. Advanced settings are subsequently discussed to further improve the algorithm for practical use. Experimental results demonstrate the advantages of the proposed algorithm over previous work. © 2013 Institute of Electrical Engineers of Japan. Published by John Wiley & Sons, Inc.