Comparing clustering algorithms performance using multiple-objective functions

Abstract

Clustering is the bunching of the data into groups of identical objects. Here each bunch is known as a cluster, each object is identical to its objects of the same cluster and different from other clusters. In this paper, we are doing an experimental study for comparing clustering algorithms using multiple-objective functions. We have investigated K-means a Partitioning-based clustering, Hierarchical clustering, Spectral clustering, Gaussian Mixture Model Clustering, and Clustering using Hidden Markov Model. The performance of these methods was compared using multiple objective functions. Multiple objectives have two core objectives: Cluster Homogeneity and separation. These multiple objective functions will be a great help to discover robust clusters in a more efficient way.