Supporting customer choice with semantic similarity search and explanation

Abstract

Semantic search and retrieval methods have a great potentiality in helping customers to make choices, since they appear to outperform traditional keyword-based approaches. In this paper, we address SemSim, a semantic search method based on the well-known information content approach. SemSim has been experimented to be effective in a defined domain, namely the tourism sector. During experimentation, one of the first requests raised from the users concerned the possibility to explain, besides the typical output of a semantic search engine, why a given result was returned. In this paper we investigate SemSim with the aim of providing the user with an explanation about the motivations behind the ranked list of returned options, with graphical representations conceived to better visualize the results of the semantic search. © 2013 Springer-Verlag.