Scholarlink

After years of development, cognitive diagnostic technology has garnered significant attention and has been extensively applied in various areas of contemporary psychological and educational measurement. In the context of smart education, cognitive diagnostic techniques can effectively identify students’ mastery of specific knowledge concepts. Although neural network-based cognitive diagnostic methods have achieved excellent results compared with traditional approaches, the relationship between students’ levels and exercise diagnostic factors has not yet been comprehensively considered, resulting in less accurate diagnostic results. On this basis, this study presents a differentiated neurocognitive diagnostic model based on students’ level, difficulty exercising and differentiation, which effectively solves the above problems; this model is referred to as DCDM. The model is structured around three key modules: the embedding module, the differentiation matching factor calculation module and the prediction module. Through the embedding module, the proficiency vector of the student's knowledge concepts, the knowledge concept association vector of the exercises, the knowledge concept difficulty vector and the differentiation vector are obtained; through the differentiation matching factor module, the relationship between the student's knowledge concept proficiency vector, the knowledge concept difficulty vector of the exercises and the differentiation vector obtained by the embedding module is calculated, and the differentiation matching factor is obtained; finally, in the prediction module, the vectors obtained before are fused with the Distinctiveness Matching Factor are fused, and then the students’ responses to the tasks are anticipated. The results of the experiment indicate that the differentiated neurocognitive diagnostic model on two publicly available datasets, ASSISTments2009–2010 and FrcSub, which is based on students’ level, difficulty exercising and differentiating improves the diagnostic effect compared with other methods.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

A Differentiated Neurocognitive Diagnostic Model Based on Student Level, Exercise Difficulty and Discrimination

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • Dongkai Qi,
  • Xiaoyu Han,
  • Zijie Li,
  • Jia Hao,
  • Jun Wang

摘要

After years of development, cognitive diagnostic technology has garnered significant attention and has been extensively applied in various areas of contemporary psychological and educational measurement. In the context of smart education, cognitive diagnostic techniques can effectively identify students’ mastery of specific knowledge concepts. Although neural network-based cognitive diagnostic methods have achieved excellent results compared with traditional approaches, the relationship between students’ levels and exercise diagnostic factors has not yet been comprehensively considered, resulting in less accurate diagnostic results. On this basis, this study presents a differentiated neurocognitive diagnostic model based on students’ level, difficulty exercising and differentiation, which effectively solves the above problems; this model is referred to as DCDM. The model is structured around three key modules: the embedding module, the differentiation matching factor calculation module and the prediction module. Through the embedding module, the proficiency vector of the student's knowledge concepts, the knowledge concept association vector of the exercises, the knowledge concept difficulty vector and the differentiation vector are obtained; through the differentiation matching factor module, the relationship between the student's knowledge concept proficiency vector, the knowledge concept difficulty vector of the exercises and the differentiation vector obtained by the embedding module is calculated, and the differentiation matching factor is obtained; finally, in the prediction module, the vectors obtained before are fused with the Distinctiveness Matching Factor are fused, and then the students’ responses to the tasks are anticipated. The results of the experiment indicate that the differentiated neurocognitive diagnostic model on two publicly available datasets, ASSISTments2009–2010 and FrcSub, which is based on students’ level, difficulty exercising and differentiating improves the diagnostic effect compared with other methods.