Exploiting Wikipedia in query understanding systems

In recent years, the free online encyclopaedia Wikipedia has become a standard resource to exploit to build knowledge base for various Natural Language Processing applications. In this paper, we exploit that resource to design a new query classification system. We explain and justify the steps we take to extract information from Wikipedia into a structured database, in order to demonstrate the validity of our design. We then show, both with mathematical reasoning and with experimental results, how to exploit the information in the database for the purpose of query classification.