|
Original research K-MEANS ALGORITHM AND SELF ORGANIZING MAPS (SOM) FOR POVERTY CLUSTERS IN MALUKU, NORTH MALUKU, PAPUA AND WEST PAPUA PROVINCESPages 71-80
Abstract
Based on data published by the Central Statistics Agency (BPS), it shows that provinces that are classified as poor are provinces in eastern Indonesia, especially Maluku, North Maluku, Papua and West Papua. In addition, based on the Presidential Decree, the provinces of Papua, West Papua, Maluku and North Maluku are also provinces that contribute most of the regencies/cities that are classified as the poorest regions. One of the classification or clustering methods is K-Means and Self Organizing Maps (SOM). SOM is an artificial neural network method that is used to group (clustering) data based on data characteristics/features. The technique used in the SOM method is done by creating a network that stores information in the form of node connections with the specified training set. The k-means clustering algorithm can be summarized by selecting the closest distance to the center and then calculating the new center based on the clustering results. This is done until there is no change in cluster members. The data used in this study is secondary data obtained from BPS publications with district/city sample units in the Provinces of Maluku, North Maluku, Papua and West Papua with a total of 63 districts/cities. This study obtained the results that by using the K-Means Cluster, 3 clusters were formed with details of cluster 1 consisting of 33 regencies/cities, Cluster 2 consisting of 15 regencies/cities and cluster 3 consisting of 15 regencies/cities. Furthermore, by using Self Orzanizing Maps (SOM) 4 clusters were formed with details of cluster 1 consisting of 45 districts/cities, Cluster 2 consisting of 2 districts/cities and cluster 2 consisting of 14 districts/cities.
Keywords: K-Means, Poverty, Self Organizing Maps (SOM).
|
Gabriella Haumahu,
PDF 