Identifying Students’ Learning Styles Based On Online Learning Behavior Analysis

With the development of mobile Internet, big data, cloud computing and other information technology, the way of receiving information has changed dramatically. The way of teaching and learning has thus undergone disruptive change. Information technology makes it possible to learn without the limit of time and space. Online learning and teaching is popular in the world.Learning management systems (LMSs) are the main medium of online education. LMSs can automate the management of distance education and support teachers to create and manage online courses and they can be used to evaluate the learning effect. LMSs aim at provide learners with individualized learning instruction without the limitation of time and space. The widely used learning management systems such as Blackboard and Moodle only focus on the management of teaching resources and provide all students with the same learning material. They hardly provide personalized services based on individual differences and have obvious shortcomings in terms of personalization and interactivity.In online learning, it is difficult for teachers to get insight into how individual students learn and behave. In order to support teachers in identifying students’learning situation and giving feedback in time, and provide learners with personalized learning resources, it requires the learning systems record and analyze the behavior of learners.At present, domestic research on online education mainly focus on the design and development of learning management systems, design of courseware and teaching management. Few studies are focus on the analysis of online learning behavior. Moreover, the studies which focus on the analysis of learning behavior collect behavior data based on filling out questionnaires. The information collected though questionnaires can only reflect the characteristics of learners’ external behavior and has a deviation between the real learning paths. The automatic approach records and collects the real data about learners’ behaviors and actions when they stay on the LMSs. It is more accurate and free from the problem of inaccurate self-conceptions of students compared to the questionnaire-based approach. Research on the analysis of learners’ online learning behavior based on the automatic approach is less.In this paper, students’ learning styles are identified automatically through the analysis of their online learning behavioral data. An improved learning style model is proposed with consideration of the social interactive characteristics. Four month’s online learning behavioral data and the offline survey data of50high school students in Beijing Digital School, a widely used LMS, are collected. An approach to detecting learners’ learning styles automatically is evaluated by the survey results and the precision can achieve to71.43%-83.67%which shows that the approach is promising for discovering learners’ learning styles.