A personalized search using a semantic distance measure in a graph-based ranking model

The goal of search personalization is to tailor search results to individual users by taking into account their profiles, which include their particular interests and preferences. As these latter are multiple and changing over time, personalization becomes effective when the search process takes into account the current user interest. This paper presents a search personalization approach that models a semantic user profile and focuses on a personalized document ranking model based on an extended graph-based distance measure. Documents and user profiles are both represented by graphs of concepts issued from predefined web ontology, namely the ODP. Personalization is then based on reordering the search results of related queries according to a graph-based document ranking model. This former is based on using a graph-based distance measure combining Minimum Common Supergraph (MCS) and maximum common subgraph (mcs) between the document and the user profile graphs. We extend this measure in order to take into account a semantic recovery at exact and approximate concept level matching. Experimental results show the effectiveness of our personalized graph-based ranking model compared to Yahoo and to different personalized ranking models performed using classical graph-based measures or vector space similarity measures.