K-Means and Fuzzy C-Means Cluster Food Nutrients for Innovative Diabetes Risk Assessment

Abstract

Packaged food and beverages often pose a risk of increasing diabetes when consumed regularly. This study aims to classify these products based on their nutritional content listed on the labels, with a focus on identifying diabetes risk. The methods employed include K-Means and Fuzzy C-Means, K-Means is used to determine initial center of cluster, while Fuzzy C-Means enhances the clustering by assigning probabilistic memberships to each data point. These methods are applied to products sold in stores in Banyumas Regency, Central Java, Indonesia. This research is the first to combine these two methods in the context of product clustering based on nutritional labels. The results indicate that packaged food and beverage products can be classified into high-risk and low-risk clusters for diabetes. Consequently, this study provides important guidance for consumers in choosing healthier.