The Modification Of The User Learning Style Model In Adaptive Learning System

With the coming of the 21st century, human is entering into the Information Age.The development of the modern information technology requests online education provide students with personalized learning mechanism, while the adaptive learning system could provide adapted courses that fits to the individual needs of the users and make the presentation of the learning content and the communication in the learning process more suitable to individual characteristics and needs, it has attracted people's attention.As the important individual tendency, learning style is commonly used as a basisi for providing adaptivity in adaptive learning system.Study abroad in this regard is relatively mature while research in our contury is relatively fragmented. Besides, the majority of the systems only use the learning style quesionnare simply as a pre-test to provide adapted courses.Due to the reasons that the questionnaire is not totally reasonable as well as learners may not fill questionnaires seriously, the result may be wrong. So how to abtain the accurate learning style to provide dynamic adaptivity is the topic of our research. As learning style model is based on traditional learning rather than online learnig, to apply it in online environments, some sort of mapping between the behaviour in traditional environments and in online environments is necessary. Based on the Felder-Silverman learning style model and relavant literature at home and abroad, this paper first discussed the relationship between user learning style and learning behavior, and based on the relavant research we can infer students' learning style according to user learning behavior.Then we distracted relavant behavior patterns from every learning style dimesion and determined behavior thresholds of every behavior pattern.Second, we used literature-based approach to calculate learning preference, proposed an architecture which modificated user learning style model according to the learning behavior, and give a corresponding modification examples. Last but not the least, this paper summarized the work conducted and looked forward to the prospects and possible issues. Meanwhile it pointed out the deficiencies and further improvement.