Preference Learning in Food Recommendation: the "Myfood" Case Study

This paper introduces MyFood, a software platform that utilizes semantic technologies and artificial intelligence to represent knowledge in the food domain and to provide smart services to its users. The platform leverages big data for large volume analysis, providing innovative, high-value services for all players in the food chain. MyFood’s design began with creating a comprehensive ontology covering food components, enabling restaurateurs and customers to consult a complete, detailed menu in compliance with the law. The platform also supports people searching for specific food services based on ingredients, verifying compatibility with ethical, religious, or food-related allergies or intolerances. MyFood’s knowledge graph allows for "intelligent" operations that increase its value for various users. This paper presents a sub-component of the MyFood platform that utilizes preference neural nets (PNN) to provide recommendations to end users based on their profile of needs related to the ingredients of the dishes. The PNN suggests the most suitable restaurant(s) for the user based on their needs. This paper provides an overview of the PNN-based recommendation system and its implementation within the MyFood platform. Overall, the MyFood platform’s preference learning component has the potential to enhance the user experience by providing tailored recommendations that match their preferences and dietary restrictions. We present insights into the early implementation and design of MyFood, highlighting its potential to transform the food industry through its cutting-edge technology and innovative services.