Traditional teaching models often fail to adjust the learning content and progress according to the actual learning situation of each student. In order to solve this problem, this paper proposes a student learning content recommendation model based on personalized exploration strategy, which can automatically push appropriate learning resources according to students’ personalized needs. By analyzing the interaction behavior of students’ historical learning data and educational videos, a student personalized knowledge tracking model is designed and combined with LinUCB algorithm to realize the recommendation of educational videos. The experimental results show that on the POJ dataset, the model improves the accuracy by 1.05% and the AUC by 2.56% compared with the traditional model. On the LLS dataset, the MSE decreased by 4.11%. The model is able to effectively capture students’ knowledge mastery status and recommend suitable learning videos based on their personalized characteristics. In addition, the model adopts parallel matrix computation with personalized exploration strategy to improve the computational efficiency and recommendation accuracy. The experimental results validate the potential of the system in the field of education, which can provide students with more personalized and intelligent learning support.
