Research On Learning Style Automatic Recognition Based On Online Behavior Data

With the improvement of the information technology,the traditional offline teaching method is not able to meet the requirements of people,online learning is becoming the mainstream trend of contemporary learning.The advent of the era of ?Internet + education? has gradually promoted the demand for online education.The first generation online education,which providing plain static learning contents,is replaced by the second generation online education guided by service innovation.Meanwhile,the second generation online education concentrates on the personalized learning,changing the online learning gravity from autonomous learning to adaptive learning.For example,the emergence of online learning system: Intelligent tutoring system and Massive Open Online Courses education platform.But the current online education system still has more or less defects.It does not consider the individual differences of learners and provides learners with the same learning materials and activities.The subject of online education can?t meet the needs of self-personality.How to integrate online education with individual differences of learners is an urgent problem.Learning style,as an important indicator to describe the individual differences of learners,which needs to collect the characteristics of learners’ behavior attributes through a large number of learning traces generated by learners during online learning,and excavate these real behavior attributes through technical means to reflect the individual characteristics of learners.Therefore,learning style recognition based on behavior data is of great significance for online adaptive learning applications.Based on the research of learning theories,this paper explores the multi-dimensional learning style and its characteristics,constructs a localized online behavior learning style model,studies and applies the methods of automatic recognition of learning style.The main research work of this study is as follows:Firstly,this paper studies the multi-dimensional learning style.Through literature research methods and content analysis methods,it studies the related theories of learning style.Combined with the characteristics of online learning in China,this article constructs a multi-dimensional learning style including three levels of physiology,psychology and society.Through the analysis of learning behavior corresponding to each dimension of learning style,it explores the classification of learning behavior,and classifies the online learning behavior data of ?The mobile self-learning classes? system with multi-dimensional learning style.Secondly,this article constructs a model of local online behavior learning style.According to the multi-dimensional learning style,compile the learning style measurement scale that conforms to localization,distribute the scale,and collect and process the data of the scale.At the same time,the online learning behavior data of a middle school ?The mobile self-learning classes? system is processed and transformed.This paper analyzes the relationship between online behavior data and learning style classification variables through single factor ANOVE test,finally develops a localized online behavior learning style model with high correlation with online behavior.Finally,the method of automatic recognition of online learning style is studied.Using the Bayesian discriminant analysis algorithm and k-means clustering analysis algorithm to mine the online behavior data and predict the learning style of learners,while comparing the results of the algorithm prediction with the results of scale checking,to detect the recognition accuracy of each dimension.Finally,this paper designs a learning style automatic recognition method which integrates the Bayesian discriminant analysis algorithm and k-means clustering analysis algorithm to improve the overall accuracy of the prediction model.This study explores a new online learning style automatic recognition method combining scale checking and data mining,which provides guidance for providing online learners with personalized online learning service,and promotes the rapid development of online adaptive learning in the era of ?Internet + education?.