两类认知诊断模型的项目参数估计及模型选择

发布时间：2018-06-29 01:20

本文选题：认知诊断模型 + DINO模型　；参考：《东北师范大学》2017年硕士论文

【摘要】：认知诊断模型是一种心理测量模型,它能为研究者提供丰富的认知信息,也可以用来评估学生的优势和劣势。尽管许多研究者越来越了解认知诊断模型,但是这个模型框架还没有被充分利用,主要有两个限制:首先,与传统的项目反应模型相比,认知诊断模型在某些情况下参数化模型是新奇的,可能更加复杂,因此,许多研究者对这些模型和它们的性质缺乏熟悉。其次,传统的项目反应模型可以通过使用商业上可用的软件进行分析,然而由于缺乏商业上可用的软件,从而阻碍了认知诊断模型的更广泛的应用。为了解决这个问题,本篇文章重点研究两类应用广泛的认知诊断模型 DINO模型和DINA模型。本文主要是采用边际最大似然估计来估计DINO模型的项目参数(DINA模型的项目参数估计已给出),并利用信息准则对DINO模型和DINA模型从返真性、普适性以及稳健性三方面进行模型的选择。最后,本文使用MATLAB软件进行了数据模拟,并对模拟的结果进行了分析。
[Abstract]:Cognitive diagnostic model is a psychological measurement model, which can provide researchers with rich cognitive information, and can also be used to evaluate the strengths and weaknesses of students. Although many researchers are increasingly aware of cognitive diagnostic models, this model framework has not been fully utilized, with two main limitations: first, compared with traditional project-response models, In some cases, the parameterized models of cognitive diagnosis are novel and may be more complex. Therefore, many researchers are not familiar with these models and their properties. Secondly, traditional project response models can be analyzed by using commercially available software. However, the lack of commercially available software hinders the wider application of cognitive diagnostic models. To solve this problem, this paper focuses on two widely used cognitive diagnosis models: Dino model and dina model. In this paper, the marginal maximum likelihood estimation is used to estimate the project parameters of the Dino model (the project parameter estimation of the Dina model has been given), and the veracity of the Dino model and the Dina model is obtained by using the information criterion. The universality and robustness of the model are selected. Finally, this paper uses MATLAB software to carry on the data simulation, and has carried on the analysis to the simulation result.
【学位授予单位】：东北师范大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：O212.1

【相似文献】