A Deep Knowledge Tracking Model Integrating Difficulty Factors
Zhuoqing Song, Sirui Huang, Ya Zhou · The 2nd International Conference on Computing and Data Science · 2021
With the rapid development of artificial intelligence education, it is of great significance to establish a scientific and effective model of learners' knowledge mastery state. Knowledge tracking (KT) as a modeling method for learners to infer knowledge level based on their learning achievements, plays a key role in intelligent education system. There are some problems in the existing knowledge tracking models, such as incomplete input features and unstable model. In order to improve the accuracy of the prediction of learners' learning effect, this paper proposes a deep knowledge tracking model (DDKT) based on the comprehensive difficulty of the topic. The model mainly includes two parts: on the one hand, the construction of the comprehensive difficulty model, on the other hand, the improvement of the model. The model combines topic difficulty with deep knowledge tracking model, enriches learners' learning process dimensions and improves the generalization ability of the model. Through comparative experiments, it is proved that the DDKT model has achieved better results.