Abstract
In recent years, there has been an increasing amount of literature on query classification. Click-through information has been shown to be a useful source for improving this task. However, far too little attention has been paid to queries in very specific domains such as art, culture and history. We propose an approach that exploits topic models built from a domain specific corpus as a mean to enrich both the query and the categories against which the query need to be classified. We take an Art Library as the case study and show that topic model enrichment improves over the enrichment via click-through considerably.