PENERAPAN CLUSTERISASI MENGGUNAKAN ALGORITMA K-MEANS UNTUK MEMBANTU SURVEILANCE MENGELOMPOKKAN STATUS GIZI PADA BALITA

Every month the Mlarak Health Center collects health data on toddlers so as to produce information related to toddler health data for each village. However, from this data, the surveillance of the Marak Health Center has not yet processed the data optimally because it still uses Microsoft Excel, where the existing data is still unstructured so that the level of health of toddlers is not yet known with certainty. Based on the existing problems, researchers created a web-based nutritional clustering application system for toddlers so that it could be a solution to simplify surveillance performance in processing data to group toddler nutrition at the Mlarak Community Health Center. Researchers used one method that is quite popular for grouping data into one or more groups, namely the K-Means algorithm with the Euclidean Distance formula and evaluating the algorithm's performance using the RapidMiner Davies Bouldin Index (DBI) method. In this study there were 2 clusters, namely cluster 0 with a diagnosis of good nutrition and cluster 1 with a diagnosis of poor nutrition. From calculating 50 data using RapidMiner, the percentage results for cluster 0 54% and cluster 1 48% were obtained with an Avg value. within centroid distance_cluster_0 is -32,529, Avg. within centroid distance_clsuter_1 is -44,866. From the results of these calculations, the smallest value is considered the best cluster, so it can be concluded that avg. within centroid distance_clsuter_0 of -32.529 produces an optimal value with an average Davies Bouldin value of -0.556.