Modeling The Self-regulated Learning Behaviors Based On Hidden Markov Models

Self-regulated learning describes the process by which students control and evaluate their learning and behavior.Self-regulating students can rely on metacognition to play an active role in learning,and promote learning outcomes by setting appropriate goals,task planning,and self-monitoring learning strategies.The ability to participate in self-regulating learning is a key factor for learning success in open online learning environment.Therefore,people who design and development an online learning environment need to consider not only the technology-related issues,but also guiding students to conduct self-regulated learning.Behavior models based on model recognition and successful learning behavior can promote personalized development of online learning systems,and then help students improve their performance in a targeted manner.Self-regulated learning is a gradual and continuous process.Although it is possible to use simple methods such as frequency and sequence analysis,exploring dynamic changes in learning strategies requires the use of algorithms that incorporate the overall student learning process.Therefore,the main work of this article is as follows:(1)Modeling and analysis of students' self-regulated learning behavior was conducted.More specifically,hidden Markov Models(HMM)were used to model student behaviors.After a comprehensive understanding and analysis of the three classic problems and corresponding algorithms of the HMM,the evaluation algorithm and re-evaluation formula were modified in the actual situation of multiple sequences.In terms of model selection,the Bayesian Information Criterion is used.To overcome the shortcomings of the Baum-Welch algorithm,such as the slow convergence speed and the difficulty of reaching the global optimal solution,the method of the K-means algorithm was introduced.(2)An online thesis writing assistance system was implemented based on self-regulated learning theoretical framework.The implementation of this system is based on Java language,adopts Spring Boot microservice architecture technology for on-demand configuration,and combines data persistence mapping tool Hibernate to realize the mapping between database and entity.Meanwhile,front-end technology such as Ajax and Thymeleaf is use to develop web pages.The purpose of the system is to provide learners with a platform for efficient paper writing in self-regulated learning environment.The system is divided into three modules:literature reading and goal setting,paper writing and opinion modification.The main functions of online thesis writing assistance are realized in the system,which makes it has a good engineering application prospect.(3)According to the self-regulated learning questionnaire,the students were divided into high and low self-regulated learning groups.A corresponding HMM model was constructed for the two groups of students' online data.In HMM,the model was interpreted as the students' learning behaviors.In order to understand students'self-regulated learning strategies,the models of the two groups were compared.The result shows that students with high self-regulated learning ability were better at self-reflection and strategy adjustment than those with low self-regulated learning ability.However,from the prospect view of the general results of the model state,the differences between the two groups of students were slight.