This study addresses grammar correction tasks in English education by proposing the GET-MF model, which optimizes correction efficiency and adaptability through modular design. By integrating unsupervised clustering technology, personalized teaching strategies are developed. Using 100 English major students from a certain university as the research subjects, the effectiveness of digital English education is validated through the Global Competence Level Questionnaire and tests. The mean global competence score of the experimental group (4.02) was higher than that of the control group (3.74), and there was a significant difference in global competence between the two groups (p=0.002). The mean global competence score of low-level students in the experimental group (3.82) was lower than that of high-level students (4.04), and there was no significant difference in global competence between low-level and high-level students (p > 0.05). Additionally, there was no significant difference in the pre- and post-test scores of global competence among low-level students (p > 0.05), while there was a significant difference in the pre- and post-test scores of global competence among high-level students (p = 0.002).
