Personalized question-answering over linked data

Abstract

A Personalized question-answering system aims to provide tailor-made answers for each user's question according to his interests and preferences. Building such a system depends essentially on creating user profiles that represent actual user interests. In this paper, we present a personalized open-domain question-answering system based on large scale ontologies. Furthermore, we describe the main components of our system and the techniques used for handling user's questions in a personalized approach based on the user profile. To evaluate our system, we introduce a user-centered methodology based on the users’ feedback. The results show a significant improvement in users’ satisfaction when their profiles are used to personalize answers.