Targeting Vaccine Hesitancy A Data-Driven Approach Using AI and Public Health Data

This study examines H1N1 and seasonal flu vaccination behaviors using machine learning models and explainable artificial intelligence (XAI) techniques. Utilizing data from the National 2009 H1N1 Influenza Survey, we developed a predictive framework employing models such as CatBoost, XGBoost, and LightGBM. CatBoost outperformed others with an accuracy of 0.696 and an F1 score of 0.688. SHAP (Shapley Additive Explanations) was used for interpretability, providing both global insights, such as the critical role of doctor recommendations, and local insights, highlighting individual decision factors. Our findings underscore the importance of addressing vaccine skepticism and improving healthcare communication to enhance vaccination uptake. These results contribute to public health strategies aimed at increasing immunization coverage and preparing for future pandemics.