Multifactor Analysis Revealing Key Factors of Chronic NCDS Based on Random Forest Models

Abstract

This article investigates chronic non-communicable diseases with a focus on lifestyle and dietary habits. By establishing separate models, eliminating the influence of highly correlated variables, and addressing data imbalance using the SMOTE algorithm, seven independent variables were identified, including basic information, lifestyle, and dietary habits. Random forest algorithm analysis revealed that smoking and alcohol consumption significantly impact the occurrence of chronic diseases, with different dietary factors associated with specific diseases. Occupational type, work intensity, and stress also have a notable influence on the risk of chronic diseases. Furthermore, increasing daily physical activity is associated with a lower disease risk. These findings contribute to a better understanding of the occurrence and management of chronic diseases, providing valuable information for prevention and treatment.