Data-driven Multiple Regression Analysis of Teaching Mode Innovation and Teaching Quality of English Education in Colleges and Universities Based on Data

For the data-driven innovation path of English education teaching mode in colleges and universities, obtaining and analyzing high-quality English teaching data is a necessary condition for realizing effective English teaching mode innovation, and this paper makes an attempt to this path. First, the teaching data are acquired and pre-processed and analyzed for correlation. Second, the improved K-means++ clustering algorithm is used to analyze the regularity between learning behaviors and teaching needs, and regress the correlation analysis results. Third, a new teaching model is constructed based on the objective laws of English learning. Finally, the feasibility of using multiple linear regression algorithm to evaluate the teaching quality of the new teaching model is explored. Attendance rate showed the highest correlation with English grades in Pearson’s correlation analysis (r=0.25), while prep rate and course grades in the multiple linear regression explained a total of 60.31% of the results of English final grades. This inspires us to focus on improving students’ listening effect, pre-preparation and course grades during the innovation process of English education and teaching mode in higher education.