Exploration and Evaluation of Fuzzy C-Means Clustering for Classification of Health Workers in West Java, Indonesia Atsila Nurtsabita, Achmad Fauzan
Statistics Department,
Faculty of Mathematics and Natural Science,
Universitas Islam Indonesia,
Indonesia
Abstract
Cluster analysis places a set of objects into several groups based on their similarity in nature or characteristics. Fuzzy C-Means (FCM) clustering divides data into members of all clusters based on their degree of membership. This study aims to obtain profiling of data on the number of health workers at community health centers (Puskesmas) in the province of West Java, Indonesia, in 2020 based on their similarity. Because the data contains multicollinearity, dimension reduction is performed on the data using Principal Component Analysis (PCA). From the results of PCA followed by FCM clustering. The goodness of the FCM method model is based on three validation values, namely: (1) Partition Entropy (PE), (2) Partition Coefficient (PC), and (3) Modified Partition Coefficient (MPC). The PE validation value was 0.339, PC was 0.793, and MPC was 0.585. Based on these three indicators, the data on health workers are divided into two clusters. Cluster 1 consists of 5 Regencies/Cities, and Cluster 2 consists of 22 Regencies/ Cities.
Keywords: Fuzzy C-Means, Health Workers, Modified Partition Coefficient, Partition Coefficient, Principal Component Analysis
