Clustering Kebutuhan Makanan untuk Meminimasi Standar Deviasi Angka Kebutuhan Gizi Menggunakan Algoritma K-Means dan K-Medoids

Abstract

To prevent nutritional problems is balanced nutrition education, which consists of the four principles of a varied diet, a clean lifestyle, physical activity and regular weight monitoring. The diversity of food at each meal is visualized as the contents of my plate which consists of one third of staple foods, one third of side dishes, and one third of vegetables and fruit. In determining balanced nutrition, the principle of Fill My Plate is used, which consists of staple foods, side dishes, vegetables and fruit. This study aims to group or cluster the Indonesian Food Ingredients Table in 2017 using the k-means and k-medoids algorithms. Data from food types that have been summarized at TKPI in 2017 are divided into 4 appropriate categories, namely fruit food clustering (Klanpenres_0), staple food clustering (Klanpenres_1), side dishes clustering (Klanpenres_2), vegetable clustering (Klanpenres_3). In the results of this study the results of clustering using the k-means algorithm that there are 109 for (Klanpenres_0), 323 for (Klanpenres_1), 166 for (Klanpenres_2), 105 for (Klanpenres_3). While the results of clustering using the k-medoids algorithm are 107 for (Klanpenres_0), 325 for (Klanpenres_1), 166  for (Klanpenres_2), 107 for (Klanpenres_3). Meanwhile, for the standard deviation results from clustering using the k-means algorithm, there are 3 categories that exceed the recommended standard for the 40 percent Nutritional Adequacy Ratio, and 1 category that is less than the standard. And the results of the standard deviation of clustering using the k-medoids algorithm, there are 4 categories that are superior to the recommended standard for the Nutrition Adequacy Rate.