Construction of learning behavior prediction model based on deep learning

With the rapid development of educational technology, it is of great significance to accurately predict students' learning behavior for improving personalized education and learning effect. This paper introduces a prediction model for learning behavior, leveraging the power of Deep Learning. Building upon the prevalent Deep FM model for click prediction, the proposed model incorporates an embedded module and a Light GBM module to bolster its feature representation capabilities and prediction accuracy. Through in-depth analysis of student behavior data on online learning platforms, including click logs, study duration and sharing, we effectively built a multi-level prediction framework. The experimental outcomes indicate that the proposed model surpasses traditional approaches in predicting learning behavior, achieving heightened accuracy and demonstrating commendable generalization capabilities.