欢迎您！东篱公司
退出
-
申报数据库
-
立项数据库
-
成果数据库
-
项目获奖数据库
中文摘要：
提出了一种既能满足属性平衡又能有效处理相似属性模式的新方法——改进的极大化整体判别认知诊断模型信息量指标法（MGCDI）．对无结构型的属性，讨论定长CD-CAT的选题策略．研究结果表明，使用相同选题策略选题时，先选择可达矩阵的所有列对应的项目比不使用这些项目的诊断准确率高．
英文摘要：
This study puts forward a new method named MGCDI, which combines modified MMGDI with CDI. It can both meet balancing attribute coverage and deal with similar attribute patterns effectively. This study attempts to construct a new item selection strategy of CD-CAT with fixed length under the unstructured hierarchy. And Monte Carlo simulation is implemented to compare the diagnostic accuracy of three item selection strategies in the bank with or without including the reachability matrix. Simulation results indicated that the new way selects the items from the reachability matrix as starting point, and then other items from the bank including the teachability matrix with maximizing MGCDI, MMGDI or CDI, the diagnostic accuracy of which is higher than selecting all items from the bank without including the reachability matrix with maximizing MGCDI, MMGDI or CDI. In future study, how to control the non-uniform exposure is worth consider. It will be an interesting topic.
同期刊论文项目
同项目期刊论文
位置：