测验Q矩阵中属性指定、选择和验证方法

Attribute Specification,Selection and Proof Methods of in the Test Q-matrix

在认知诊断评估中,构建正确测验Q矩阵十分关键但比较困难。本文在简要介绍认知诊断评估之后,从认知诊断模型视角,重点阐述和评价了知名模型或框架下(规则空间模型、属性层级模型、融合模型、回归树、确定性输入噪音与门模型、证据中心设计和认知设计系统)测验Q矩阵中属性指定、选择和验证的方法,指出了各方法的优势和劣势。最后,归纳了测验Q矩阵构建的一般性步骤和方法,并讨论了未来可能研究方向。

In cognitive diagnostic assessment, to build up the right test Q - matrix is crucial but diffi- cult. From the perspective of the cognitive diagnostic model, the paper elaborates and evaluates at- tribute specification, selection and proof techniques in the test Q - matrix under the famous models or frameworks（ Rule Space model, attribute hierarchy model ,fusion model, return tree, deterministic in- put noisy AND -gate model, evidence center design and cognitive design system）, points out the ad- vantages and disadvantages of each method, summarizes the general steps and methods of the test Q -matrix construction, and discusses possible future research directions.