Aplikasi Segmentasi Jenis Makanan Berdasarkan Kandungan Gizi Menggunakan Algoritma K-Means

Abstract

Public understanding of food nutrition remains limited due to the lack of accessible educational tools. This study aims to develop a web-based food segmentation system capable of clustering food items based on their nutritional characteristics using the K-Means algorithm. The research employs the Food Composition Dataset, focusing on key attributes including nitrogen factor, fat factor, and specific gravity. The methodology consists of data preprocessing, determining the optimal number of clusters using the Elbow Method, and visualizing the clustering outcomes through Principal Component Analysis (PCA). The results indicate that the optimal number of clusters is K = 3, with clear separation demonstrated by PCA, which shows an explained variance of 82%. The resulting clusters represent groups of foods with similar nutritional profiles, such as high-protein, high-fat, or high-density items. The clustering results were implemented into an interactive Streamlit-based web application, allowing users to explore and interpret the segmentation results easily. The study concludes that the K-Means algorithm is effective for grouping foods based on nutritional attributes, and the developed system can serve as a practical tool for nutritional education and balanced diet analysis.