Comparison of K-Means and Fuzzy C-Means Algorithms for Clustering Based on Happiness Index Components Across Provinces in Indonesia

Abstract

Cluster analysis is a multivariate technique aimed at grouping objects into several clusters based on the characteristics they possess. This study aims to determine the clustering results of 34 provinces in Indonesia based on the indicators of the happiness index for the year 2021 by comparing non-hierarchical cluster analysis methods, namely K-Means and Fuzzy C-Means. K-Means is a non-hierarchical cluster analysis that divides objects into cluster groups based on the distance of objects to the nearest cluster center, while Fuzzy C-Means is a cluster analysis that uses a fuzzy grouping model where data becomes a member of a cluster formed based on membership degrees ranging from 0 to 1. Based on the research results, it is known that clustering with both K-Means and Fuzzy C-Means methods forms three clusters. Based on the standard deviation values between groups and the standard deviation ratio, the best method is the Fuzzy C-Means method because it has a larger standard deviation between groups and a smaller ratio compared to the K-Means method, which is 0.6680004. Therefore, this study concludes that the Fuzzy C-Means method is more optimal than the K-Means method.