Construction method of personalized learning feature model based on knowledge graph

Abstract

A construction method of a personalized learning feature model based on a knowledge graph comprises the steps: constructing a course-oriented knowledge graph, wherein entities correspond to knowledgepoints and learning resources; describing a relationship between knowledge point entities by adopting a directed acyclic graph; describing and learning a relationship between resource entities by utilizing a Bloom cognitive hierarchy; constructing a student personalized cognitive model, and adopting transfer learning to solve the problem of scarcity of personalized information of a single student;adopting active learning to solve the value evaluation problem of the selected data; and constructing a dynamically evolved student personalized cognitive model by utilizing a Markov chain theory. The invention discloses a construction method of a personalized learning feature model based on a knowledge graph. By comprehensively utilizing the knowledge graph, the graph theory, the Markov chain, the Bloom cognitive hierarchy theory and other methods, the personalized learning characteristics of the students are effectively positioned, the accuracy of positioning the personalized characteristics of the students is improved, and therefore quantitative decision support is provided for improving the learning efficiency and the teaching effect.

Description

technical field [0001] The present invention relates to the field of computer and education technology, and in particular to a method for constructing a personalized learning feature model based on a knowledge graph. Background technique [0002] Personalized features are the basis of personalized learning and an important basis for improving teaching efficiency. The methods of constructing student models based on personalized learning characteristics mainly include: modeling methods based on knowledge tracking, modeling methods based on cognitive diagnosis, and modeling methods based on machine learning. [0003] The modeling method based on Knowledge Tracing (KT) was originally proposed by Corbett and Anderson. They regard students' mastery of knowledge points as hidden variables, and infer the probability distribution of hidden variables based on the correct and wrong results of students' answers to exercises. . The advantage of the KT model is that it can model the ind...

AI Technical Summary

Problems solved by technology

[0009] However, the above-mentioned methods of analyzing students' personalized cognition based on cognitive diagnosis and machine learning usually only consider a few factors related to cognition

Images