认知诊断自适应测验选题策略探新

The Exploration of Item Selection Strategy of Computerized Adaptive Testing for Cognitive Diagnosis

提出了一种新的认知诊断自适应测验选题策略和题库按项目所包含的属性模式分层方法.与传统方法相比,该方法不仅提高了测量精度而且可以提升选题速度.

A new item selection strategy of computerized adaptive testing for cognitive diagnosis, and the method of partition of the item pool according to the attribute pattern in the item, called as pattern of pattern-stratified method are proposed. The new strategy improves the measurement accuracy and enhances speed of item selection, compared with a traditional method of Xu et al＇s algorithm based on Shannon＇s Entropy.