COMPARATIVE PERFORMANCE OF RANDOM FOREST, SUPPORT VECTOR MACHINES, AND NEURAL NETWORKS IN PREDICTING NUTRITIONAL QUALITY OF FOOD PRODUCTS

Abstract

ABSTRACT: In this research, we perform a comparative analysis of Random Forest (RF), Support Vector Machines (SVM), and Neural Networks (NN) for predicting the nutritional quality of food products. Accurate prediction of nutritional quality is vital for enhancing consumer health and supporting informed dietary choices. Using a dataset comprising key nutritional attributes such as protein, fat, and carbohydrate content, we evaluate the performance of these machine learning models in both classification and regression tasks. Metrics such as accuracy, precision, recall, F1-score, and Root Mean Square Error (RMSE) are employed to assess the models. Our results show that Neural Networks outperform both RF and SVM in terms of accuracy (94.2%), precision (93.0%), and recall (94.5%), while also achieving the lowest RMSE (0.039) for continuous nutritional score prediction. RF also demonstrated competitive performance, while SVM lagged behind in both classification and regression. This study provides insights into the applicability of these models for food quality prediction, with implications for the food industry and consumer health monitoring.