Student modeling with timed assessment information

This paper presents a Bayesian belief network based approach to modeling students with their timed learning assessment information. Our modeling method updates beliefs using the assessment result of a student's learning performance on a learning object along with the time spent on the assessment. The modeling process requires few parameters and takes linear time, reducing the high computational cost usually associated with knowledge acquisition and updating student models. Our student modeling system can model the performance of a student on a topic more accurately by taking the time spent on the assessment, in addition to the student's slips and guesses in the assessment, into account.