A New Approach to Recommend XML Facet-Values by Fusion of the Statistic Feature and User’s Cognitive Model

Abstract

Faceted search is a kind of exploratory information query technique, which is a complement to keyword search. To recommend the most valuable facets, the existing facet-value recommendations are based mainly on the experience of domain experts or on the statistic features. While recommending same facet-values and results to different users, these methods may not meet different user’s need. This paper proposes a new facet recommendation method in XML faceted search based on both the user's cognitive model and statistic feature of facet-values. First, we propose and build a user's cognitive model, then we give the definition of fusional facet recommendation score and the fusional facet-value recommendation score, based on these definitions, we propose a new facet recommendation method. The experiment results in searching through Internet Movie Database show that our method is more effective than the method based either on the statistic feature of facet-values or on users' cognitive model respectively.