Smart Recipe Recommendation for a Healthy Diet Based on Nutrient Density and Macronutrient Balance

Since the time of Hippocrates, healthy nutrition has been considered an important key to maintaining health, with the aim of reducing illness and prolonging life as much as possible. In these endeavors, Information and Communication Technologies (ICT) and Artificial Intelligence (AI) can be of great use. This paper presents an experimental web system for recommending recipes based on the principles of healthy eating. The system, named Recipe Smart Picker, was developed based on the theory of nutrient density and nutrient profile models. The idea is to choose specific ingredients and then to recommend recipes from the recipe database that both include those ingredients and best satisfy the nutritional balance, both micronutrient and macronutrient. Nutrient density principle compares content of beneficial nutrients in the dish in proportion to their Recommended Daily Values (RDV), as well as non-beneficial nutrients in proportion to their Maximum Daily Values (MDV). The goal of macronutrient balance is to have protein, carbohydrates and fat as close as possible to their recommended relative proportions. The system was developed with the goal of promoting healthy eating, but also as a platform for student scientific work in the fields of ICT and AI application in nutrition and gastronomy.