Optimal machine learning models and factors for global infectious diseases: An ecological analysis

Area of Science:

• Public Health
• Epidemiology
• Machine Learning

Background:

• Infectious diseases pose a significant global health threat, exacerbated by socioeconomic and environmental shifts.
• Effective predictive modeling and identification of key factors are essential for disease prevention and control.
• Standardized data and reusable models are needed for infectious disease research.

Purpose of the Study:

• To establish an optimal machine learning model adaptation system for infectious diseases.
• To provide standardized data support for infectious disease research.
• To offer reusable model references for future studies.

Main Methods:

• Utilized infectious disease incidence data from Global Burden of Disease 2021 and socioeconomic/air pollution indicators from Our World in Data.
• Evaluated eight machine learning models, including Gradient Boosting Decision Tree (GBDT), Light Gradient Boosting Machine (LightGBM), and Random Forest (RF).
• Employed SHapley Additive exPlanations (SHAP) to quantify factor contributions globally and regionally.

Main Results:

• GBDT excelled in sexually transmitted and respiratory diseases; RF was optimal for all infectious diseases; LightGBM for enteric; GBDT for neglected tropical diseases.
• Key identified factors included child dependency ratio, unsafe drinking water, particulate matter 2.5 exposure, and socio-demographic index.
• Model performance varied by disease type, with R² values ranging from 0.434 to 0.854.

Conclusions:

• Adapted machine learning models effectively identify critical factors influencing infectious diseases.
• Core strategies for reducing the global infectious disease burden involve improving water safety, clean energy access, diet, and child health services.