Implementation of the Healthy Eating Habits Support System Based on User Taste Preferences and Nutritional Balance

This paper proposes a healthy eating habits support system that recommends menu selections based on a user’s taste preferences and the requirements of long term nutritional balance. This system is comprises a nutritional management system (NMS) and a Kansei retrieval system (KRS). NMS adopts a tabu search method to generate a large number of nutritionally balanced menus by combining multiple recipes from a recipe database and stores these menus in a “candidate list.” From this candidate list, KRS retrieves menus that are compatible with a user’s taste preferences and presents these menus to the user. KRS utilizes Kansei retrieval agents that represent a user’s taste preferences and impressions that, subject to a user’s evaluation of the presented menus, evolve based on a hybrid model composed of a genetic algorithm and simulated evolution. A simulation using 7,260 actual recipes incorporating 13 types of nutrients demonstrated that the system presented a large number of menus nutritionally balanced over the long term, and predicted a user’s taste preferences with more than 80% accuracy after continuous use for eight weeks.