Learning Path Recommendation of Intelligent Education Based on Cognitive Diagnosis

Many learning path recommendation methods of intelligent education have been proposed and implemented. However, many of them have problems or limitations, which may result in unsatisfactory recommendation results. Therefore, this research aimed to study the learning path recommendation method of intelligent education based on cognitive diagnosis. Combined with a cognitive diagnostic model (CDM), personalized and accurate learning paths were recommended to students. This study fully considered the multidimensional features of interaction between students and knowledge when designing the CDM, described the cognitive process, and provided a comprehensive ability modeling method based on cognitive rules. A neural matrix decomposition model was constructed, which incorporated the personality features of students’ comprehensive ability level based on cognitive rules, thus obtaining their predicted scores in various knowledge and skills learned. The model consisted of three parts, namely, the generalized matrix decomposition part, the multi-layer perceptron part and the NeuMF layer. Finally, the experimental results verified that the constructed model was effective.