SWITCHtoHEALTHY AI-Based Family Nutrition Recommendation System: Promoting the Mediterranean Diet

Area of Science:

• Nutrition Science
• Artificial Intelligence
• Family Health

Background:

• Modern families struggle with healthy, sustainable diets due to time constraints.
• The Mediterranean diet (MD) offers health and environmental benefits but is challenging to implement.
• Existing systems lack personalized, family-focused MD planning.

Purpose of the Study:

• To present and validate the SWITCHtoHEALTHY AI-based Family Nutrition Recommendation System.
• To generate personalized meal plans aligned with Mediterranean diet guidelines for families.
• To assess the system's accuracy and adherence to nutritional rules.

Main Methods:

• Developed two AI recommendation engines: one for adults and one for children.
• Utilized an expert-validated Mediterranean food dataset and MD-based nutritional rules.
• Validated systems with real-world family intervention data.

Main Results:

• Achieved 90% accuracy in generating family meal plans meeting MD dietary rules.
• The adult recommender exceeded 90% accuracy in calorie and nutrient content estimation.
• Demonstrated successful integration of school menu options for children's plans.

Conclusions:

• AI-based recommendation systems show potential for promoting healthier, sustainable family diets.
• The SWITCHtoHEALTHY system effectively generates personalized, nutritionally balanced MD meal plans.
• AI can help overcome barriers to consistent MD adherence in households.