Study on the Individual and Environmental Factors of College Adaptation of Freshmen: Research Based on Machine Learning

Abstract

In this study, we used the method of multiple regression and machine learning (Ridge regression and Lasso regression) to investigate the influence of individual and environmental factors on school adaptation at different time. The results shows that: (1) College students' adaptation is negatively correlated with shyness, loneliness, Internet addiction and non-adaptive cognition; meanwhile positively correlated with self-awareness, active coping style, self-esteem, interpersonal trust and social support. (2) Lasso regression of machine learning can effectively reduce dimensions and the number of variables, further simplify the model to strengthen the explanation. (3) The key factors of college students' adaptation are shyness, coping style, self-esteem, loneliness, Internet addiction and social support.