Mining Heterogeneous Individual Factors in Sustained Compliance With Major Epidemic Prevention and Control Policies

In major epidemic prevention and control, sustained public compliance with policies is crucial. This study is based on three-phase survey data from 480 residents and integrating machine learning with the SHAP framework, analyzes the heterogeneity and dynamic mechanisms of public compliance behaviour during long-term control. The findings indicate that public opinion intensity, frequency of preventive education, and perceived certainty of sanctions are key driving factors, with significant phase heterogeneity and nonlinear interaction effects observed. The research breaks through the limitations of static analysis, providing theoretical and empirical support for precise and dynamic epidemic control policies.