多分属性层级结构下引入逻辑约束的理想掌握模式

发布时间：2018-06-14 06:40

  本文选题：认知诊断 + 多分认知属性　； 参考：《江西师范大学学报(自然科学版)》2017年03期

【摘要】：多分属性比传统的2分属性提供更多更详细的诊断反馈信息,具有广阔的应用前景.在多分属性情境下,当属性之间存在层级结构时,会出现原2分属性情境下不存在的逻辑问题:如果被试仅低程度地掌握了父属性,那么他是否还有可能高程度地掌握子属性?从逻辑上讲,这种"父属性掌握程度低而子属性掌握程度高"的发展情况并不具有普适性.对此,该文首先在多分属性情境下,基于现有的计算理想掌握模式的方法提出了满足"属性掌握水平约束假设"的理想掌握模式计算方法.然后,通过模拟研究说明该逻辑约束的使用方法及忽略该逻辑约束可能对诊断结果带来的危害.
[Abstract]:Multi-division attributes provide more detailed diagnostic feedback information than traditional two-component attributes, so it has a broad application prospect. In the case of multiple attributes, when there is a hierarchy between the attributes, there will be a logical problem that does not exist in the context of the original two-score attribute: if the subject only grasps the parent attribute to a low degree, is it possible for him to master the child attribute to a high degree? Logically, this kind of development of "low degree of master of parent attribute and high degree of mastery of subattribute" is not universal. In this paper, firstly, based on the existing methods of computing ideal mastery model, an ideal mastering mode calculation method is proposed, which satisfies the assumption of "attribute mastering horizontal constraint" in the context of multi-partition attribute. Then, the method of using the logic constraint and the harm of neglecting the logic constraint to the diagnosis result are illustrated by simulation research.
【作者单位】： 北京师范大学中国基础教育质量监测协同创新中心;江西师范大学计算机信息工程学院;浙江师范大学教师教育学院心理系;
【基金】：全国教育科学规划教育部重点课题(DBA150236) 国家自然科学基金(31360237,31500909,31300876,31160203,31100756,30860084,11401271)资助项目
【分类号】：B842.1
，

本文编号：2016498