AI-Powered Recommendation System: A Web-based Approach for Personalized Nutrition and Dietary Goal Management

Abstract

The increased rate of nutrition and lifestyle-related diseases has caused an urgent need for intelligent and personalised nutrition management systems. The traditional approaches to dietary recommendations are primarily based on universal meal planning guidelines without addressing specific health needs, requirements, preferences, fitness objectives, or restrictions on access to foods. As a result, in this paper, a web-based personalized nutrition recommendation framework based on Artificial Intelligence (AI) is proposed in order to overcome these drawbacks. The user health and nutrition datasets are first preprocessed by Robust Scaler (RS) to reduce the effect of outliers and to normalize the features. Then, the Autoencoder learns deep dietary features and dietary patterns. Moreover, Select from Model (SFM) selects the most important nutritional features to be optimized so as to achieve a higher recommendation efficiency and a higher accuracy in the prediction. Finally, the XGBoost model is deployed in the AI recommendation engine where it classifies dietary goals and provides personalized meal recommendations, sending them to the web dashboard, chatbot, and ordering module. The experimental results demonstrate the superiority of accuracy, precision, recall, F1-score, recommendation reliability and long-term dietary adherence when compared to a traditional approach.