Analysis of Web-based Student Learning Behaviour and Teaching Adjustment Strategies in a Blended Teaching Model

Abstract

In the context of the rapid development of Internet technology and the economy, the blended learning model has been supported by the environment, and web-based learning has begun to spread widely. In this paper, after using the Z-score data standardisation method to process the students’ online learning behaviour data under the blended learning mode, the correlation between each learning behaviour characteristic and performance is calculated by the correlation analysis method. Then, the SMOTE-XGBoost-FM model is proposed to predict students’ academic performance based on behavioural characteristics, and the performance prediction results are used to assist teachers in making teaching adjustments and interventions and obtain good teaching results. It was verified that the correlation between homework scores and academic performance among the online learning behavioural features was the highest (0.91), and the accuracy, precision and recall of the fusion model’s performance prediction were all above 0.95. It has also been found that after analyzing learning behavior and adjusting teaching strategies, the frequency of students’ online behavior significantly increases, and they can actively access various resources on the platform during the learning process. This paper explores the mechanism and strategy of differentiated teaching adjustment based on online learning analysis, which achieves accurate diagnosis, multi-dimensional follow-up and effective adjustment and intervention for students by making full use of behavioural analysis to diagnose and analyse students.