Machine Learning Framework for Dietary Pattern Discovery Using UK Survey Data

A new explainable machine learning framework for discovering dietary patterns from the UK National Diet and Nutrition Survey (NDNS) has been proposed in a preprint on arXiv. The framework uses an unsupervised-to-supervised approach to identify reproducible and interpretable dietary patterns. Adult participants aged 19 and above from NDNS Years 12-15 were analyzed using 25 energy-adjusted nutrient and food-group features. Three clustering algorithms—K-means, Gaussian Mixture Models, and Agglomerative Clustering—were compared across k values from 2 to 8, with stability and dietetic interpretability prioritized alongside internal validation metrics. The selected K-means solution with k=4 identified four distinct dietary patterns: high fat/meat and sodium; higher fibre fruit-vegetable micronutrient; high free-sugar snacks and sugary drinks; and dairy/cereal calcium-rich saturated-fat. The framework aims to translate complex dietary data into actionable counseling priorities for clinical settings.