Exercise Recommendation Method Combining NeuralCD and NeuMF Models

Exercises are an indispensable part of the student's learning process. Students choosing appropriate exercises can improve learning efficiency and speed up learning progress. However, due to the explosive increase of exercise resources, students face difficulties in choosing suitable exercises. Therefore, this article proposes an exercise recommendation method combining NeuralCD (Neural Cognitive Diagnosis) and NeuMF (Neural Matrix Factorization Model) models, called NeuralCD-NeuMF method. First, this method uses the NeuralCD model to obtain students' knowledge point proficiency and exercises' knowledge point difficulty, outputing students' exercises mastery level. Then combined with the NeuMF model, the NeuMF incorporated students' Personality Characteristics (NeuMF-PC) model is constructed to predict students' scores on exercises. Finally, according to the difficulty level, suitable exercises are recommended to each student. The effectiveness of the NeuralCD-NeuMF method is verified through a large number of experiments.