A New Classification Method Based On CRM In Cognitive Diagnosis

As the development of social, people have not satisfied with the achievement thattest only gives an evaluation of the macro-level, but hope to get more informationabout the examinees’ cognitive states from the examinees’ response in order to makemore objective assessment and remedy for individual examine.Cognitive diagnosisdevelops fast as the education, psychological, mathematics and computer developing,especially of the pattern recognition,artificial intelligence. Rule Space Model (RSM)and Attribute Hierarchy Model(AHM) is two common kind of Cognitive DiagnosticModel(CDM). Based on the two cognitive diagnosis methods: Sun Jianan proposeda new cognitive diagnosis method named Generalized Distance Discrimination. Basedon the Q matrix theory, GDD can discriminate an examiner’s attribute mastery patternby computing generalized distance between his observed response pattern and eachideal response pattern.In this paper, an extension of the GDD for polytomously-scored items (hereafterreferred to as the Poly-GDD) is proposed. In simulation study, a comparison ofclassification accuracy is made among GDD and AHM based on Graded ResponseModel. The results of simulation showed that GRM-GDD is better than GRM-AHM.The trends of the classification rate varied with the slip are consistent at all kinds ofattribute hierarchies.