Explainable artificial intelligence for studying the impact of environmental factors on health

The interplay between environmental factors and human health has gained significant attention in recent years, particularly in the context of noncommunicable diseases (NCDs) such as cardiovascular diseases, diabetes, and respiratory illnesses. This thesis investigates the role of the exposome—encompassing the totality of environmental exposures over a lifetime—in shaping NCD risks and outcomes. Advanced machine learning (ML) and Explainable Artificial Intelligence (XAI) methods are employed to analyze complex, multidimensional datasets from environmental and health domains, uncovering critical relationships and patterns. The research includes three studies: a regression model predicting respiratory cancer mortality with XAI insights into feature importance; a classification model distinguishing provinces with high respiratory cancer mortality rates; and regression approaches for predicting mortality from endocrine, nutritional, and metabolic diseases, highlighting key environmental and socioeconomic predictors. By integrating predictive modeling with XAI, the work provides actionable, transparent insights for researchers, policymakers, and healthcare practitioners, fostering evidence-based interventions and policies to address public health challenges in Italy.