Research on Cognitive Diagnostic Model Based on BP Neural Network

Abstract

This paper proposed a cognitive diagnostic model based on BP neural network and maximum likelihood function, with an ultimate goal of realizing the cognitive diagnostic computerized adaptive test. The model was verified through the actual data from 1029 British middle school students who participated in the English reading test of the Program for International Student Assessment (PISA). The experimental results show that both the model’s mode accuracy rate and the average attribute accuracy rate are relatively ideal. Moreover, the model, which is of flexibility, can be effectively applied to cognitive diagnosis in small-scale classrooms.