Personalizing nutrition and recipe recommendation using attention mechanism with an ensemble model

Nutrient management in the context of this proposed work aims to quantize the consumption of essential nutrients in an efficient format such that it leads to a healthy and balanced lifestyle. This paper presents an intelligent nutrition management and recipe recommendation system tailored to individuals’ nutritional profiles, using an ensemble model augmented by an attention mechanism. The system quantifies user nutritional deficiencies based on blood analysis and personal preferences, generating personalized food and recipe suggestions to address these gaps. By integrating multiple supervised learning algorithms such as Random Forest, XGBoost, and MLP, the model dynamically prioritizes nutrients relevant to the user’s needs. Leveraging data from the National Institute of Nutrition, recipes are recommended in video format, aiming to enhance users’ health and dietary habits. The proposed model outperforms baseline systems in detecting nutritional deficiencies and offers efficient, personalized recipe recommendations through a user-friendly web and mobile interface.