A Temporal Model of Learner Behaviors in OELEs using Process Mining

Abstract

Open-ended learning environments (OELEs) present learners with complex problems and a set of tools for solving these problems. Developing logging mechanisms that capture learners’ interactions with the system provide a wealth of trace data that can be employed for studying relations between their behaviors and performance. Such analyses provide a framework for making the OELE intelligent in that it can adapt its feedback to meet the needs of individual learners. In our previous research, we have developed learner modeling schemes that are based on sequential pattern mining (SPM) and Hidden Markov models (HMMs) to represent and track the temporal sequence of learners’ interactions with the OELE. We briefly discuss the pros and cons of these models, and then propose a process modeling approach to capture the temporal nature of learners’ behaviors. We apply the process modeling method to data collected from students working with the Betty’ Brain OELE, where students learn about scientific processes by building causal models.