Research On Automatic Recognition Method Of Learning Style In Online Courses

In recent years,the vigorous implementation of the policies related to education informatization has made major universities pay close attention to the deep integration of education and information technology.At the same time,internet technology is constantly innovating and gradually becoming industrialized.The concept of"Internet +"has also been proposed and gradually applied to various fields.With the support of this policy and technology,online education is booming,and it became the mainstream learning methods in the field of primary and secondary schools,colleges and universities and even correspondence education.However,the lack of monitoring and management of online education platforms and services,and the lack of attention to the individualized development of learners in online education models are also reasons for restricting their in-depth development.Therefore,for network educators,how to combine the network education and the individualized characteristics of learners is an urgent problem to be solved.When learners conduct online learning,they will produce a series of"learning traces".By analyzing these learning behaviors to obtain the relevant characteristics of learners,and thus guiding learners'personalized learning,which has important promotion significance for the in-depth development of online education.Based on the idea of data mining,this study takes the semester course data of a class in Central China Normal University as the research object,and uses the decision tree to construct the'Felder-Silverman learning style prediction model'.The main work of this study is divided into the following points:(1)Analysis of the characteristics of online learning behaviors that affect learning styles.Analyze the various dimensions of the Felder-Silverman learning style model to determine the relationship between learning style and learning behavior.Establish an online learning behavior classification to clarify the online learning behaviors that can be collected.Combine the characteristics of the online learning behavior platform and curriculum,analyze and extract relevant behavioral characteristics that can affect the learning style.(2)Data preparation and attribute correlation analysis of learning style prediction model in online courses.This study used the scale of Soloman&Felder to obtain learning styles of learners,collected the learner's behavior data in the platform of'cloud class',and a series of operations such as data conversion and cleaning were performed to obtain a standardized online behavior data set.Finally,we use the correlation analysis method to analyze the relationship between each dimension learning style and various specific online learning behavior characteristic.(3)Research and application of automatic learning style recognition methods in online courses.The decision tree algorithm is used to construct the learning style model,and the post-prune is performed during the training process.The decision tree model is analyzed.It is found that the behavioral attribute attributes determined by pruning to predict the decision tree model of each dimension are consistent with the results of correlation analysis.The online learning behavior-learning style model is proposed.At the same time,the learning style model constructed in this study is compared with the accuracy of the model of the representative research,and the difference is analyzed.And relevant suggestions for online education development are provided.In this study,we have developed a learning style automatic recognition method in online courses by combining data mining and questionnaire survey,and provided relevant suggestions for learners'personalized online learning service push,which provided a certain guiding significance for the diversified development of online education.