Penerapan Principal Component Analysis untuk Reduksi Variabel pada Algoritma K-Means Clustering

Abstrak

K-Means clustering is a widely used clustering algorithm. However, it has the disadvantage that the performance of clustering data decreases if the variables of the processed data are immense. The complex variables problem in K-Means can be overcome by combining the Principal Component Analysis (PCA) variable reduction method. This study uses seven indicator variables for the welfare of the people of West Java Province in 2021 to measure the welfare level of districts/cities. The results of the analysis obtained two principal components based on eigenvalues. Clustering from cluster analysis with the K-Means with variable reduction using PCA formed the three best clusters where the number of members of each cluster consisted of 12, 8, and 7 districts/cities.