Enhancing Dietary Guidance with Machine Learning: A Stacking Classifier Approach to Personalized Nutrition

Abstract

Personalized nutrition is increasingly important for promoting better health by customizing dietary recommendations to fit individual needs. Traditional diet planning methods often struggle with inefficiency and lack the flexibility to adapt to changing personal preferences. This paper addresses these challenges by introducing a machine learning-based solution that utilizes a stacking classifier model. The system focuses on key health indicators such as BMI (Body Mass Index), BMR (Basal Metabolic Rate), and TDEE (Total Daily Energy Expenditure) to provide accurate and tailored nutrition advice. By integrating models like Random Forest and Long Short-Term Memory (LSTM), the approach not only improves accuracy but also enables real-time adaptability. The findings demonstrate that machine learning, particularly the stacking classifier, serves as a highly effective tool for enhancing the accuracy and relevance of personalized dietary recommendations. The model achieved 95.5% accuracy, 93% precision, 97% recall, 94% F1-score, and 97% ROC AUC score, thereby highlighting its strong predictive performance across key metrics.