Control variables.

Although much research has been done on occupational burnout, most studies focus on employees (EM), with insufficient attention given to the self-employed (SE). This study adopts a novel quantitative approach using methods like the entropy weight method, multiple linear regression, and the Boruta algorithm to analyze burnout differences between EM and SE. Results indicate that EM are more influenced by stress and external, deficiency-induced burnout, while SE are driven by interest and internal, imbalance-induced burnout. Social network support is a key factor in predicting burnout for both groups. Interestingly, increased financial support from family raises burnout risk for EM, while greater social network support increases burnout risk for SE. These findings reflect inherent differences in work patterns and resource mobilization strategies. This research expands the application of the Job Demands-Resources model and provides insights for tailored mental health interventions across different occupational contexts.