A new model developed to predict mental resilience in college students could have significant implications for how universities address the growing mental health challenges facing this demographic.
The research, published in the International Journal of Information and Communication Technology, used statistical methods to investigate how students cope with stress and adversity. Fulian Liu of the Mental Health Education Center for College Students at Wuxi Institute of Technology in Wuxi, China, suggests that the study could help educational establishments identify those at risk of mental health struggles and by anticipating challenges, offer appropriate interventions to preclude severe psychological distress in the most vulnerable.
Mental resilience, or toughness, might be defined as one’s ability to cope with stress and setbacks. It is closely related to psychological resilience and well-being. College students, often navigating a complex mix of academic pressures, social challenges, and personal transitions, can be particularly vulnerable to mental health problems. This, in turn, can have a detrimental impact on their academic performance and their quality of life during their studies and afterwards. An understanding and of mental toughness and the ability to foresee problems arising before they become serious could be an important part of supporting students.
Current models for tracking and studying mental health have limitations. They often struggle with what is referred to as “overfitting” wherein the model performs well on training data but fails to work well with novel information. Additionally, irrelevant or redundant variables can cloud or colour the predictive process, reducing the reliability of the model. The new work tackles these problems and uses an optimized version of Elastic Network Regression (ENR) to balance multiple predictive factors statistically speaking. This approach uses a Bayesian optimization algorithm (BOENR) to fine tune the parameters within the model based on prior knowledge and is able to improve accuracy by avoiding the overfitting problem.
The new model has a much lower prediction error than other models, offering an accuracy rate of above 94 percent and outperforming five commonly used models in terms of predictive reliability.
Liu, F. (2024) ‘Mental toughness prediction model of college students based on optimal elastic network regression‘, Int. J. Information and Communication Technology, Vol. 25, No. 10, pp.19-33.