The Research Of Grouping Algorithm Of Users Of Chinese Platform Based On Personal Information By BIRCH Clustering

With the rapid development of Internet technology, online teaching becomes themain form of Chinese language teaching. Online learning, which is fast, efficient, andunrestricted, is a useful complement and extension to traditional teaching model.Meanwhile, in order to make the original abstract content more meaty,online teachingcan take advantage of images, animations, videos and other multimedia tools. It cansignificantly improve the efficiency of Chinese language teaching on the basis oflearners'interest.However, due to geographical reasons, the less interaction between teachers andstudents leads to that most of online teaching platforms cannot modify their teachingscheme considering learners'own qualifications. As a personalized teaching platform,it should be able to dividend learners into different groups due to their personaldemands,learning abilities and styles.In the northeast base construction project of Hanban. This project is belong to thedevelopment of northeast base information platformnine of nine key sub-topics ofbase project. Because of the imperfectness of Chinese language teaching system, westart to analyse a method to personalize learners. This paper studies Chinese learnersgrouping method based on registration information. The main research results of thispaper can be outlined as follows:1. Put forward a basic information model of learnersPuts forward a basic information model applied for grouping learners'registration information by studying personalized learning theory and the modelbuilding. This model is a kind of cognitive modeland and it can reflect learners'needs, abilities and individualized learning styles.2. Pretreat learners heterogeneous information The basic information model includes mutual heterogeneous data. In order torealize the learners grouping,this paper preprocesses every learners'heterogeneousdata and make variables have equal weight. Futher more, this paper defines thedistance between the basic information models and the distance between the learnersgrouping.3. Achieve clustering based on the basic information of learnersThe basic information of Chinese learners mainly comes from the registeredinformation. BIRCH algorithm based on hierarchy clustering is applied. By creating acluster feature tree, we finally get the corresponding grouping results.We set the registration data of 136 test students as training data, and search it bychanging the branching factor B and maximum clustering several M.Then we applyeffectiveness index evaluation and satisfaction questionnaire evaluation to thegrouping strategy we get. For the following massive learners grouping,we provide acertain reference.The main innovations of this paper is outlined as follows:1. Put forward a basic information model applied for grouping of learners'registration information. Chinese learners basic information model includes age, sex,personality traits, cultural background, Chinese level, learning motivation and style,which can reflect their ability of learning.2. Implement grouping algorithm of Chinese learners base on registeredinformation. Because of heterogeneous learners registration information, I putclustering method as research and application object, with researching in this paper.The clustering features structure(CF) of original algorithm is changed into SCF(SCFstructure consists of the number of data points in clusters, cluster of mold andtwo-dimensional matrix which is storaged classification attribute value frequency), Itis used to calculate similarity of heterogeneous learners model.This paper groups learners according to the basic information by BIRCHalgorithm and realize to group learners based on their different learning needs,abilities, personalized style, and some other factors. Then strategies that suit fordifferent kind learners will be recommended by experts on Chinese. Not only can learners get the right method to study in advance, but also it can embody the teachingcore with stylization and personalization.Because of project requirements, currently this paper only groups based onlearners registration information and Project has been planned that learners'feedbackinformation will be added later to realize clustering in the learning process.