An interpretable polytomous cognitive diagnosis framework for predicting examinee performance

Xiaoyu Li, Shaoyang Guo, Jin Wu, Chanjin Zheng

Abstract: Highlights•We proposed the Polytomous Cognitive Diagnosis Framework to re-encode polytomous data.•PCDF used Cumulative Category Response Function theory to keep model interpretability.•PCDF can seamlessly integrate into existing dichotomous cognitive diagnostic models.•Our experiments focused on model interpretability and more scoring categories.