The Research Of Learning Resource Recommendation Based On Student Attributes

The online learning platforms have been widely used with the background of educational informatization.Along with the growth,the data generated by the learners' online learning process makes the learners face the predicament of information overload and information trekking.How to rapidly select suitable and accurate learning resources in massive resources is a huge challenge.The research on online social networks has become gradually mature,but in the field of education,the concept of clustering has not yet been extensively explored and researched.In this paper,the research of student attributes in domain social learning network is intended to reveal students' behavior activities from the social learning network online learning system,which to the group recommendation of learning resources.Based on student online learning,we research students' characteristics,then use students' learning target and learning efficiency as clustering parameters for community detection.The purpose of constructing the student attribute community is to analyze the differences in student behavior and learn effects in specific student groups made up by different attributes.The method is used to recommend resources for student groups,which can make up the waste of resources that caused by previous individual resource-recommendation forms,and it is helpful to solve problems such as cold-start and data sparsity.Based on the student attributes,this paper studies the resource recommendation method in the community with the same or similar attributes students.Main content as follows:First,to obtain student attributes.By extracting keywords from the 9 activities in the Moodle platform,collecting data generated by the students during the use of the platform and preprocessing the data;collecting relevant data of three types of research variables: student attributes,emotional status,and being involved in or out of learning state,building a three-tier model of student attributes.Second,multi-community and its clustering mining.By studying the characteristics of student attributes and combining clustering algorithms to process data,the student groups with similar attributes are divided into a community structure,multiple student attribute communities are discovered,and student conditions in the same community are studied.Based on a systematic analysis of the current learning situation of a student community,objective learning factors such as the students' group cognitive level and mastery of knowledge points in the same community,and subjective learning factors such as learning preferences and acceptance levels are derived.Third,according to the characteristics of the student's attribute community,this paper use the collaborative filtering recommendation algorithm to propose a recommendation mechanism that fuses student community and learning objectives.In the resource recommendation process,we proceed from two aspects.On the one hand,we establish the relationship between learning resources,mainly through the semantic correlation analysis of learning resources to connect the learning resources in the teaching process.On the other hand,we establish the relation between students and learning resources,analyzing the student communities separately,realizing the learning situation and preference information of students in the same community,and then to recommend learning resources.