A Comparative Study Of K-Medoids And Fuzzy K-Means Clustering For The Selection Of Optimal Cloud Service Provider

Abstract

Cloud providers offer various services, including storage, computing power, and data management, making selecting the best provider for a particular application challenging. Clustering algorithms can simplify the process by grouping cloud service providers based on cost, reliability, security, and performance similarities. K-medoids and fuzzy K-means clustering are two widely used clustering algorithms that can be applied to this problem. K-medoids is a non-parametric algorithm that selects representative points, called medoids, for each cluster. At the same time, fuzzy K-means is a soft clustering algorithm that assigns each data point to multiple clusters with varying degrees of membership. By grouping cloud service providers using these algorithms, comparing and selecting the best provider for a particular application is more accessible. This paper discusses the application of K-medoids and fuzzy K-means clustering to select the best cloud service provider and highlights the advantages and limitations of each approach