From "Events" to "Activities": Creating Abstraction Techniques for Mining Students' Model-Based Inquiry Processes

In this paper, we present a technique that we have developed to transform sequences of technical events into more abstract actions and semantic activities. The sequences of more abstract units are then used for discovering patterns of students’ interaction with computer models using heuristic miner. Our proposed approach automatically segments sequences of technical events that occurred during model runs and pauses and, on the basis of the nature of technical events that occurred during model runs and pauses, clusters them into actions. Then, using heuristic rules, it classifies actions into activities. We demonstrate the usefulness of our multilevel abstraction for extracting and exploring characteristic patterns of students’ interaction with computer models. Our study shows that each abstraction level could help to identify distinct characteristics of students’ interaction.