Mining and Analysis of Self-regulated Learning Process Model: Based on Hidden Markov Model

In online learning, mining self-regulated learning based on clickstream data has gradually attracted attention, but mining and comparing self-regulated learning behavior patterns of learners with different achievements has been paid little attention. This study uses hidden Markov model to identify the self-regulation behavior patterns of three learning achievement groups in the learning management system. The results show the self-regulation process model of the three groups is mainly a two-way transition between perception and control. Mastery learners aim at knowledge gain and adjust their learning through continual monitoring behavior. Goal- oriented learners mainly complete learning objectives, assign more time to sense learning information, adjust learning performance through evaluation behavior, and have less learning monitoring behavior. The self-regulation process model of baseline learners has the most connections, like that of mastery learners. However, this group has the least learning behaviors, and mainly through adjusting online evaluation behaviors to improve learning performance. In sum, hidden Markov model can identify self-regulated learning behavior patterns of learners with different achievements. The research provides practical support for mining self-regulated learning mechanism, and has theoretical and methodological significance for the research and development of self-regulated learning dynamics.