Abstract—This study is based on a proposed framework that integrates a co-occurrence graph and ontology to develop an adaptive reinforcement recommendation system, referred to as Co-occurrence and Ontology via Reinforcement Learning (CoO-RL). The implementation of CoO-RL on healthy recipes and ingredients facilitates a recommendation system for ingredient substitution within recipe data. This document elucidates how CoO-RL executes its algorithm, providing implementation details to clarify the innovative problem-solving approach utilized by the framework. The application’s contribution, stemming from the efficacy of the food recommendation system, effectively meets its objective by suggesting ingredient substitutions based on user feedback, user profile, and constraints. Conceptually, this work offers an alternative method for generating recommendations, wherein the dataset encapsulates similarity relationships among data instances, structured as an ontology network. The research articulates and substantiates its intent at both the application and design levels. By employing co-occurrence and ontology techniques to analyze and adjust food ingredients according to user preferences, the results demonstrate an accuracy of 80% in the recommendations, while ensuring that the proposed menu ingredients maintain appropriate nutritional value. Consequently, this research effectively promotes and enhances overall nutritional food choices in alignment with nutritional principles.
 
Keywords—healthy food, co-occurrence, ontology, reinforcement learning, recommendation system


Copyright © 2025 by the authors. This is an open access article distributed under the Creative Commons Attribution License which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited (CC BY 4.0).