Research On Learning Material Recommendation System Of Learning Material For Adaptive Learning Based On Cognitive Diagnosis

Traditional tests tend to report only one test score without considering the potential psychological processes and cognitive characteristics of the participants.Compared to traditional tests,cognitive diagnosis can evaluate information such as individual cognitive process or knowledge structure.At present,most cognitive diagnostic tests are limited to diagnose students' attribute mastery.Cognitive diagnostic assessment does not provide answers to questions such as how students learn after diagnosis and how teachers select appropriate learning materials.The current education is mainly class teaching,and the monotonous teaching process may make students lose motivation for learning,which to some extent restricts the development of students.With the development of the Internet,adaptive learning,which provides students with different learning resources based on their differences,has become a research focus of scholars.Therefore,it is necessary to carry out the study of adaptive learning based on cognitive diagnosis.According to the previous research,this paper believes that the adaptive learning system should include four models: material model;attribute model;learner model;material recommendation model.The material recommendation model is the focus of the research.This study divides the problem scenarios into two types: each material contains only one attribute;each material can contain multiple attributes.In different problem scenarios,different penalty functions are established according to the same nine recommendation rules,and genetic algorithm is used to select appropriate learning materials for each learner.In this paper,an adaptive material recommendation algorithm based on cognitive diagnosis is investigated through two experiments.The research results show that:(1)The adaptive material recommendation algorithm based on cognitive diagnosis developed in the study has a relatively ideal effect and is basically feasible.This algorithm can recommended corresponding learning materials to different learners?(2)Learning materials selected by genetic algorithm have lower value of penalty function.Compared with random algorithm,learning materials selected by genetic algorithm are more suitable for learners.(3)Learning materials selected by genetic algorithm have a high success rate in cognitive features(attributes,difficulty,mastery probability and hierarchy)and noncognitive features(media,content,background and time).(4)In the study of Tatsuoka's fractional subtraction data,it is found that the adaptive material recommendation algorithm based on cognitive diagnosis has ideal effects and wide potential application scenarios.