The Study Of Classification Accuracy In Cognitive Diagnosis

Traditional test only provides a total score and computer adaptive test (CAT) not only provides examinee's score but also estimates examinee's ability. However, both two methods have not studied students'knowledge states. Cognitive diagnosis, combined with cognitive psychology, metrology and pedagogical research is gradually developed into an emerging science and demonstrates its importance and potential in education. Cognitive diagnostic assessment (CDA) is designed to measure specific knowledge structures and processing skills in students so as to provide information about their cognitive strengths and weaknesses. Teachers in the theory of cognitive diagnosis can be guided to design a test, which can identify students with cognitive errors.Rule space model (RSM) belongs to a branch of statistical pattern recognition and classification problems in statistics. This approach has two phases: First is the feature variables selection phase, and second is the statistical pattern classification phase. In the feature selection phase, we determine feature variables that in RSM are knowledge and cognitive or thinking skills called attributes, and then classification groups as knowledge states defined by which attributes a student can or cannot do. The core of pattern recognition is pattern classification and classification accuracy is an important study of cognitive diagnosis.A good diagnostic test is one that not only estimates individuals'overall ability levels and distinguishes individuals'cognitive state, but also provides opportunities to observe the process of student responses and increase the amount of information available from student answers. According to incomplete statistics, there are as many as sixty kinds of cognitive diagnostic models (CDMs) in foreign countries. Deterministic inputs noisy"and"gate (DINA) model is introduced in the paper. To improve the classification accuracy, a new selection strategy, which combines the reachability matrix with cognitive diagnostic information index is developed. Monte Carlo simulation shows that the pattern classification rate and mean marginal classification rate are improved.