Ontology-based dietary recommendation system for Chinese children and adolescents: development and a pilot validation study.

Abstract

Dietary imbalances among Chinese children and adolescents are prevalent, but most recommender systems are adult-oriented and not adapted to Chinese dietary patterns or school-family meal contexts. This study aims to developed and evaluated an ontology-based dietary recommendation system for Chinese children and adolescents to improve their dietary status. An ontology knowledge base was constructed from dietary guidelines, expert knowledge, and food and meal data. A multi-criteria decision framework generated feasible meal plans, optimized portion sizes with a genetic algorithm, and selected solutions that promote food diversity. We conducted a preliminary evaluation of the system in two pilot settings: (1) a group-level experiment using 30 days of school lunch data from a primary school, and (2) an individual-level experiment using 30 days of dietary records and health check-up data from 30 middle school students. Diet quality was assessed using the Chinese Children Dietary Index (CCDI), nutrient adequacy rates, and food-group diversity. In the group experiment, system-recommended lunches achieved a significantly higher average CCDI score (117.39 ± 6.61) than actual school lunches (103.41 ± 9.44, p < 0.001). In the individual experiment, recommended full-day meals scored higher in CCDI (110.91 ± 9.70) compared with self-selected diets (90.61 ± 10.77, p < 0.001). Across both settings, the system reduced deficiencies in vegetables, fruits, and aquatic products with consistent improvements in nutrient adequacy and food diversity, while maintaining alignment with dietary guidelines. To the best of our knowledge, this study presents the first ontology-driven dietary recommendation system tailored for Chinese children and adolescents. By integrating structured knowledge representation with advanced decision-making algorithms, the system demonstrates promising improvements in dietary quality, nutritional balance, and food diversity across both school and family scenarios. These findings highlight the system's potential as a practical tool for promoting healthy eating behaviors, and informing nutrition education.