User Profiling for Query Refinement Submission Type: Completed Research Paper

Query refinement is a famous technique which helps users to issue more effective queries. However, current query refinement techniques reply the same candidate query list to users who issue the same query without considering their diverse search interests. In this paper, we propose a novel framework of user profiling -- modeling the user’s interests at the task level for query refinement to improve the ranking of the refined queries, and thus, ultimately improve the ranking of the relevant search results. Our approach incorporates the LDA model to extract topic-based user’s interests from search sessions and search tasks, a four-descriptor to learn users’ long-term and short-term interests, and a module for task identification and update. Experimental results show that our task-based user profiling method contributes to an increased precision, and it produces more accurate alternative queries.