Building User Interest Profiles Using DBpedia in a Question Answering System

In this paper, we explore the idea of building an adaptive user interest model. Our proposed system uses implicit data extracted from a user’s search queries to select categorical information from DBpedia. By combining the categorical information collected from multiple queries and exploiting the semantic relationships between these categories, it becomes possible for our system to build a model of the user's interests. This model is designed to be responsive to changes in the user's interests over time by including concepts of aging and expiration. Our system also includes mechanisms to pinpoint the correct categories when an ambiguous term is queried. We evaluated our system using a predefined set of test queries and shown to correctly model user short term and long term interests.