A comprehensive survey of traditional, merge-split and evolutionary approaches proposed for determination of cluster number

Today's data mostly does not include the knowledge of cluster number. Therefore, it is not possible to use conventional clustering approaches to partition today's data, i.e., it is necessary to use the approaches that automatically determine the cluster number or cluster structure. Althoug...

Full description

Saved in:
Bibliographic Details
Published in:Swarm and evolutionary computation 2017-02, Vol.32, p.49-67
Main Authors: Hancer, Emrah, Karaboga, Dervis
Format: Article
Language:English
Subjects:
Automatic cluster evolution
Clustering
Evolutionary algorithms
Knee point
Validity indexes
Citations: Items that this one cites
Items that cite this one
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:Today's data mostly does not include the knowledge of cluster number. Therefore, it is not possible to use conventional clustering approaches to partition today's data, i.e., it is necessary to use the approaches that automatically determine the cluster number or cluster structure. Although there has been a considerable attempt to analyze and categorize clustering algorithms, it is difficult to find a survey paper in the literature that has thoroughly focused on the determination of cluster number. This significant issue motivates us to introduce concepts and review methods related to automatic cluster evolution from a theoretical perspective in this study.
ISSN:2210-6502
DOI:10.1016/j.swevo.2016.06.004