Interpreting eye movements with process models

Though eye movements provide a wealth of information about how humans interact with computers, the analysis of eye movement data can be extremely tedious and timeconsuming. This paper outlines an automated approach to tracing eye movements, that is, interpreting eye movement protocols based on an underlying process model. The proposed tracing methods utilize techniques such as hidden Markov models to relate observed eye movement protocols to the predictions of the process model. These methods have been applied successfully in the domain of equation solving and will be extended to several other task domains.