Comparing two classification methods based on the attribute hierarchy method and the DINA model

Cognitive diagnostic assessment (CDA) for education can be recognized as an effective data mining approach in education research. In CDA, one of the critical steps is how to classify correctly the examinee in the most probable of the attribute mastery patterns or classes. The article mainly discussed and compared two classification methods: method A based on the attribute hierarchy method and the maximum a posteriori estimation of the DINA model. The Monte Carlo simulation results showed that the DINA model outperforms the AHM in terms of the correct classification rate for the entire patterns; however, for some patterns with larger test information conditioned on associated abilities, AHM performs better. Further, a rule of combining the results of two methods was proposed. Thus, one can exploit their individual advantages in order to reach an overall better performance than could be achieved by using each of them separately.