The development of education informatization has given rise to a large amount of learning behavior data, which provides new ideas for education management and teaching quality improvement. This paper constructs a learning behavior computational analysis model based on the XGBoost algorithm and explores the strategies for improving the quality of college English teaching. Starting from students’ English learning behavior data, the study extracts six features: speaking practice, English writing, English classroom homework, consulting English dictionary, memorizing English words, and English listening practice, and establishes a computational analysis model through data preprocessing and feature engineering. The results show that the XGBoost algorithm performs well in the computational analysis of learning behaviors, with an accuracy of 0.9592, a recall of 0.9644, and an F1 value of 0.9618, which is significantly higher than that of traditional machine learning methods. Teaching experiment validation shows that the teaching strategy formulated based on the results of computational analysis can effectively improve the quality of teaching, and the average value of students’ English performance in the experimental group improves from 62.33 points in the pre-test to 88.08 points in the post-test, which is significantly higher than that of the control group, which is 63.81 points. The post-test questionnaire showed that the strategy use level of students in the experimental group increased from “low-moderate” to “high” before the intervention. The study proposes teaching strategies such as constructing an ecological classroom, implementing behavioral preventive measures, and creating an English teaching environment, which provide theoretical basis and practical guidance for improving the quality of college English teaching.
