ABSTRACT
College students across the Chinese mainland have begun returning to school, but they are under 15-day medical observation. It is worth paying attention to whether college students can adjust their mind and rationally cope with the influence of COVID-19 in a collective atmosphere to start their normal study again. This paper aims to establish a predictive diagnostic model of automatic adaptability of 947 medical college students after returning to school with the personal characteristics, socio-demographic, self-evaluation, and collective atmosphere. A data set containing 757 medical students was evaluated with 10-fold cross-validation using seven classifiers, and a data set from another 190 medical students were tested for effectiveness. The CatBoost method had the highest prediction accuracy of 96.32% and the AUC of 80.0%. Machine learning technology can establish a generalized prediction model and discover the latent disease risks caused by the impact of social behavior events, to help managers and psychologists take reasonable countermeasures in advance.
