A Relevance-Focused Search Application for Personalised Ranking Model

Abstract

The assumption that users' profiles can be exploited by employing their implicit feedback for query expansion through a conceptual search to index documents has been proven in previous research. Several successful approaches leading to an improvement in the accuracy of personalised search results have been proposed. This paper extends existing approaches and combines the keyword-based and semantic-based features in order to provide further evidence of relevance-focused search application for Personalised Ranking Model (PRM). A description of the hybridisation of these approaches is provided and various issues arising in the context of computing the similarity between users' profiles are discussed. As compared to any traditional search system, the superiority of our approach lies in pushing significantly relevant documents to the top of the ranked lists. The results were empirically confirmed through human subjects who conducted several real-life Web searches.