A Cognitive Diagnosis Method Based On Mahalanobis Distance

Cognitive Diagnosis is the core of a new generation of educational theory test,and diagnostic classification is an important part of cognitive diagnosis. Cognitive diagnosis model is a tool to achieve the cognitive diagnostic.Cognitive diagnosis model is established to estimate the individuals and to obtain more information about each individuals.On the basis of Generalized Distance Discrimination(GDD) and Hamming Distance Discrimination(HDD), this paper makes the experiments that used the same experimental conditions, try to study application of Mahalanobis Distance to discriminate individuals’ cognitive state. Using entropy as weighted of Hamming distance between ORP and IRP, then according to the principle of minimum distance discriminant classification. Under the same condition form experimental results, For the Linear and Convergent attribute hierarchy, the classification accuracy of new method is slightly higher than that of HDD R method, HDD B method and GDD;For Unstructured,the new method is better than the other three methods; For Divergent, the new method is the same as the HDD B, HDD R, but better than that of the GDD; For Independent, the new method is slightly less than HDD R, HDD B,but better than that of the GDD. Compared to GRM – GDD, the new method has better effect of the classification when use polytomous scoring model.