Content-based Recommendations within a QA System using the Hierarchical Structure of a Domain-Specific Taxonomy

In this paper we present several content-based recommendation methods for a QA system that rely and use extensively the structure of a domain-specific taxonomy. Our goal is to add semantics to a typical content-based RS in order to improve the quality of the recommendations by mapping relevant keywords from the existing taxonomy to the available questions. In order to test and evaluate the effectiveness of the above mentioned methods, we conducted a supervised survey where we asked several users to rate the recommendations delivered using these methods. The results show that by combining the results retrieved by these methods, we obtain a range of recommendations that satisfy a variety of user expectations.