ANALISIS k-MEDOIDS DENGAN VALIDASI INDEKS PADA IPM DAERAH 3T DI INDONESIA

Human development is a paradigm that places humans as the main target of all activities, namely controling over resources, improving health and education. The Development Index (HDI) in Indonesia varies each district, especially 3T areas. area an classified underdeveloped, remote outermost terms economy, health, education infrastructure. k-Medoids method partitional clustering for grouping several objects into clusters. This algorithm uses medoid center cluster, so it robust to data containing outliers. study aims classify regions based on find out which areas require more attention from government optimizing numbers. size object similarity calculated by using Euclidean distance Manhattan distance, selection best number clusters, internal cluster validation, such Calinski – Harabasz index, Gamma Index, Silhouette index. result this showed were four measurement, having index score 37.15764, 0.7821181, 0.3354435.