Abstract：In this paper， an SHT method， a Q-matrix revision strategy based on the DINA model is proposed innovatively. The study employs Monte Carlo simulation techniques to conduct an in-depth comparison with existing similar approaches， comprehensively evaluating its feasibility and accuracy. This method has the following three advantages. 1) Efficient revision and robustness validation： the SHT method demonstrates remarkable revision efficacy under varying levels of error rates， significantly enhancing the precision of the Q matrix， thereby validating its robustness.2) Dual benefits in small and large sample scenarios： compared with domestic and international counterparts， the SHT method particularly excels in small sample scenarios， with its robustness and performance advantages becoming more pronounced when confronted with high error rates.3) Distinct advantages with complex datasets： in empirical data analysis， the SHT method not only enhances the fitting capacity of cognitive diagnostic models but also exhibits more conspicuous advantages over other algorithms when dealing with datasets of high dimensionality complexity and relatively limited samples.
李 波,胡誉骞,章 勇,田 怡. 基于认知诊断DINA模型的Q矩阵优化：一种结合样本筛选与假设检验的新策略[J]. 华中师范大学学报(自然科学版), 2025, 59(1): 111-124.
LI Bo,HU Yuqian,ZHANG Yong,TIAN Yi. Optimization of the Q-matrix in cognitive diagnostic modeling base on DINA： a new approach combining sample selection and hypothesis testing. journal1, 2025, 59(1): 111-124.
2025, Vol. 59 
