Using Machine Learning to Analyze the Predictors of Life Satisfaction: Focus on Lifestyle Attitudes and Psychological Factors

OBJECTIVES: Life satisfaction is an essential indicator of quality of life, and enhancing it can contribute to individual well-being strategies. Because it is a complex concept, a comprehensive approach is needed to address it effectively. Machine learning offers a unique statistical opportunity to address this challenge effectively. In this study, we examined how lifestyle parameters, psychological issues, and psychological processes predict life satisfaction. METHODS: The study included 1366 participants, representing the general population. Lifestyle factors were self-reported, and included exercise frequency, alcohol consumption, smoking, body mass index, and regularity of social rhythms. The participants also completed several assessment scales, such as the Life Satisfaction Scale, the Hospital Anxiety and Depression Scale, the Acceptance and Action Questionnaire-II, the Tuckman Procrastination Scale, the Big Three Perfectionism Scale-Short Form, and the Brief Social Rhythm Scale. Machine-learning methods were used to evaluate the statistical parameters, with root mean square error values of 3.9, 3.6, and 3.7 for gradient boosting, extreme gradient boosting, and light gradient-boosting machine, respectively. RESULTS: The top five factors influencing life satisfaction were identified as depression scores, psychological inflexibility, marital status, social rhythm, and procrastination. Psychological inflexibility influences the impact of depression on life satisfaction. Factors that are difficult or impossible to change, such as age, gender, education, and chronic disease, ranked lower on the list. By contrast, psychological and environmental factors that can be improved had strong predictive power. CONCLUSIONS: These findings offer opportunities for enhancing life satisfaction and underscore the responsibility to address these factors.