Advancing food equity through explainable AI: identifying place-based factors and conditions of food security

Food behaviors and food security are shaped by complex socioeconomic factors with significant implications for public health, economic stability, and social equity. Understanding these relationships is essential for developing effective policies and interventions that promote sustainable food systems and advance food equity.

This study applies explainable artificial intelligence (XAI) techniques to identify key features influencing household food behaviors. Findings from the XAI analysis are integrated with inverse reinforcement learning (IRL) to model and examine expert behaviors associated with achieving food satisfaction and improved food security outcomes.

The XAI analysis identified household health conditions, spending patterns, and frequency of store visits as primary drivers of food behaviors and preferences. The IRL modeling further revealed behavioral patterns and decision-making strategies associated with higher levels of food security and dietary satisfaction.

These findings highlight actionable pathways for improving food security by identifying the conditions and behaviors that support equitable food access. Integrating XAI and IRL provides a novel approach for translating complex data into policy-relevant insights, offering guidance for interventions aimed at fostering sustainable, health-promoting, and equitable food systems.