Food Diet Recommendation System Using AI

Abstract

This collection of scholarly papers reviews diverse interventions for dietary and health management, employing methodologies ranging from technology-driven systems to clinical and policy-based research. Several studies focus on integrating artificial intelligence (AI) and machine learning (ML) to deliver personalized dietary recommendations. One notable study described the “Diet Engine” system, which achieved 86% accuracy in food identification using image recognition and a deep learning network with 295 layers. It also reported that the Extra Tree Classifier reached 99% accuracy in predicting diet plans. Another study applied the K-means clustering algorithm to group users for personalized dietary planning, obtaining 96% accuracy. Clinical studies explored dietary interventions in specific populations. For example, one investigation tested the MIND diet on individuals with multiple sclerosis and found increased antioxidant status and weight loss in females. Another clinical trial targeting participants with cardiovascular disease reported modest improvements in diet quality after intervention-based education, although lifestyle modifications did not significantly change cardiovascular risk factors. A related study confirmed the feasibility of low-energy diets for individuals with obesity and chronic kidney disease, showing notable reductions in both weight and blood pressure. Beyond individual health, broader socioeconomic and environmental concerns were also addressed. One policy-focused paper examined the “affordability gap” of nutritionally adequate diets across countries such as the Dominican Republic, Ethiopia, and Indonesia, recommending this metric to guide social assistance and nutrition programs. Another study raised concerns about the health risks associated with plastic additives in food, emphasizing that infants are particularly vulnerable due to exposure levels exceeding safety thresholds.Finally, an analysis of diet and lifestyle mobile applications found that many incorporated behavior change techniques correlated with app quality. However, the review found limited scientific validation and a lack of safety features in most applications, underscoring the need for more evidence-based, user-protective digital health tools.