Do User (Browse and Click) Sessions Relate to Their Questions in a Domain-specific Collection?

We seek to improve information retrieval in a domain-specific col-lection by clustering user sessions as recorded in a click log and then classifying later user sessions in real-time. As a preliminary step, we explore, the main as-sumption of this approach: whether user sessions in such a site relate to the question that they are answering. The contribution of this paper is the evalua-tion of the suitability of common machine learning metrics (measuring the dis-tance between two sessions) to distinguish sessions of users searching for the answer to same or different questions. We found that sessions for people an-swering the same question are significantly different that those answering dif-ferent questions, but results are dependent on the distance metric used. We ex-plain why some distance metrics performed better than others.