Analysis of the sleep quality of college students from different knowledge areas using a data mining approach

Abstract

Sleep is an essential physiological process involved in memory consolidation, metabolic and endocrine homeostasis, and immune system regulation. Therefore, good sleep quality is vital for maintaining physiological homeostasis. Poor sleep quality is prevalent among college students and affects both physical and mental health. Conventional statistical methods, such as logistic regression, are commonly used to generate predictive models of sleep quality and have been extensively applied to Health Sciences students, but their use has been less studied among students from other disciplines, such as Engineering and Exact Sciences. Data mining can help overcome certain limitations of these conventional methods, such as multicollinearity, by uncovering associations that might otherwise have gone unnoticed. In this study, we separately analyzed two samples of students from Health Sciences and Engineering and Exact Sciences. We found significant correlations between sleep quality and attributes such as perceived sleep quality, sleep latency, sleep duration, drug use, and the use of medication for depression and anxiety. Decision trees identified different predictive attributes between the two samples. These findings offer a novel insight into sleep quality among college students and may support informed decision-making and targeted interventions.