Research On Computer Educational Data Mining Based On Cognitive Diagnosis Methods

Under the environment of modern curriculum and the innovation of educational method,cognitive diagnosis is an important research in educational data mining.And it plays a pivotal role in it.Cognitive diagnosis refers to inferring cognitive state of students by analyzing their performance on a set of questions.This method can provide teachers with the potential learning state of students.At the same time,computer science,as an emerging discipline,has its unique char-acteristics.Although traditional cognitive diagnosis methods can play a role in com-puter science courses,they lack the diagnosis of practical ability,which makes them unsuitable for computer science teaching in some aspects.In addition,traditional cog-nitive diagnosis models ignore the connections of knowledge concepts in most cases,which leads to insufficient use of this information.As a result,traditional models cannot be effectively applied to some actual teaching scenarios.Due to the above problems,this paper will use educational data mining methods to provide auxiliary information for computer science teaching from following three aspects.·Construction of cognitive diagnosis model for computer science teaching.Cog-nitive diagnosis is introduced into computer science teaching by adding students'programming ability and modeling the students' theoretical and experimental abilities at the same time.It can be applied to computer science courses and provide students' cognitive information needed in computer teaching.Experi-mental results show that our model can accurately predict students' performance in computer science courses and provide informative cognitive reports.·Construction of the prerequisite relationship of knowledge concepts.The mas-sive information contained in Wikipedia is used to extract features of knowledge concepts.Combined with the active learning method,fewer artificial labels are used to extract knowledge concepts prerequisite relationship.It lays a strong foundation for introducing the relationship of knowledge concepts into cognitive diagnosis.Experiments show that our method can obtain an accurate prerequisite relational classifier and reduce the labels required for model training.·The prerequisite relationship of knowledge concepts is introduced into the cogni-tive diagnosis models.By using the prerequisite relationship of knowledge con-cepts,the existing cognitive diagnosis models are improved to make full use of the information in teaching process.And the influence of prerequisite relationships of knowledge concepts on different cognitive diagnosis models will be explored.Experiments on a variety of cognitive diagnosis models show that the prerequisite relationship of knowledge concepts can improve the performance of continuous cognitive diagnosis models effectively.In summary,this paper constructs the prerequisite relationship of knowledge con-cepts,and improves cognitive diagnosis methods based on the characteristics of com-puter science teaching,to provide comprehensive auxiliary information such as stu-dents'cognitive status for computer course teaching.This information can help teach-ers understand students' knowledge mastery and practical ability,so as to formulate and improve teaching plans.