A Diet Recommendation System for Persons with Special Dietary Requirements

Abstract

A food recommender system assists consumers in selecting daily diets based on specific dietary guidelines. This study outlines the unique nutritional requirements of foods and the purposes of a diet recommender system to assist the person who is the victim of various diseases due to deficiency or abundance of Nutrition. The recommender system utilizes two inputs, the nutrition-based food dataset, and the the person's health profile. In this research, knowledge from the physiology field describes reasons for diseases that grow up due to inadequate or excessive intake of Nutrition in food. The diet recommendation system uses the K-means clustering algorithm, the food inference algorithm, and the patient nutrition calculation algorithm to prescribe the best food per individual patient's nutrition requirement. The food inference algorithm is based on two parameters: The K means clustered food dataset and the patient nutrition calculation value. Our study utilizes the confusion matrix to evaluate the system's performance to obtain Precision, Recall, and F measure. The result of our research is a diet recommender system that recommends the food dish(s) having an appropriate quantity of Nutrition, taking into account the gender, age, and current nutrition status of the person. This system's significant advantage is reducing the excessive usage of medicine and preventing side effects. The diet recommender system can also help a person to pre-cure from disease and get rid of appointments with a physician for medication and food prescriptions.