KLASTERISASI DESTINASI WISATA BERDASARKAN KOORDINAT GPS DENGAN METODE KERNEL K-MEANS

Abstract

Although K-Means Clustering, a sort of non-hierarchical cluster analysis, is frequently employed, it has limitations when processing data with non-linearly separable characteristics (no distinct boundaries) and overlapping clusters, or when visually comparing the results of clusters to those of other clusters. When dealing with data that have overlapping and non-linearly separable cluster features, the Gaussian Kernel function in Kernel K-Means Clustering can be applied. The input data must be displayed using kernel functions in a new dimension, which is where Kernel K-Means Clustering and K-Means diverge.
 The purpose of this study is to determine the concept of steps and the results of Kernel K-Means Clustering analysis to classify tourist destinations in the Pesonajawa database based on GPS coordinates. The data used was obtained from the PesonaJava.com database, which consists of 28 tourist destinations located on the island of Java. The silhouette score is used in this study to evaluate the cluster results.
 The evaluation of the cluster results based on the research results revealed that the best number of clusters was K = 6, with a sigma value of = 5.