Modeling Search Behavior in Heterogenous Environments

A large part of today’s information retrieval studies focuses on modeling user search behavior to improve search algorithms, evaluate their quality and simulate users. Most of previous research on user modeling considers homogeneous search environments, where search devices, users, their search tasks and corresponding search results are assumed to be uniform. However, real-world search environments are highly heterogeneous. In this position paper, we elaborate on heterogeneous search environments and outline challenges in user modeling that arise in this case, e.g., the availability of rich user feedback and the need of device-specific and cross-device models. We also set up an agenda for research on user modeling in heterogeneous search environments, which includes acquisition of user interaction data, development of advanced user models and application-oriented evaluation of these models.