BACKGROUND
Knowledge graph-based food recommendations are critical for the nutritional support of older adults. However, the integration of disease-based nutritional principles and preference-related requirements needs to be optimized in current recommendation processes.
OBJECTIVE
This study aims to design a knowledge graph-based personalized meal recommendation system for community-dwelling elders, and to conduct the preliminary effectiveness testing.
METHODS
ElCombo, a personalized meal recommendation system, was developed driven by user profiles and food knowledge graph. User profiles were established from a survey of 96 community-dwelling elders. Food knowledge graphs were supported by data from websites of Chinese cuisine recipes and eating history, consisting of five entity classes (dishes, ingredients, category of ingredients, nutrients, and diseases), corresponding attributes, and relations between entities. A personalized meal recommendation algorithm was then developed to generate packaged meals as outputs, considering disease-related nutritional constraints and personal dietary preferences.
RESULTS
Our recommendation system, ElCombo, was evaluated by comparing the dietary diversity and diet quality of its recommended meals with those of the autonomous choices of community-dwelling elders. Simulation experiments based on retrospective data of 96 community-dwelling elders revealed that recommended meals had significantly higher diet quality and dietary diversity (P<0.001). Two representative cases involving community-dwelling elders with and without eating history demonstrated the recommendation system’s potential to fulfill complex nutritional requirements associated with multiple morbidities.
CONCLUSIONS
ElCombo proved superior performance in simulations compared to autonomous choices, implying the potential for improving dietary practices for community-dwelling elders. Future studies are needed to optimize its real-world application and refine data handling abilities.