K-means properties on six clustering benchmark datasets

This paper has two contributions. First, we introduce a clustering basic benchmark. Second, we study the performance of k-means using this benchmark. Specifically, we measure how the performance depends on four factors: (1) overlap of clusters, (2) number of clusters, (3) dimensionality, and (4) unb...

Full description

Saved in:
Bibliographic Details
Published inApplied intelligence (Dordrecht, Netherlands) Vol. 48; no. 12; pp. 4743 - 4759
Main Authors Fränti, Pasi, Sieranoja, Sami
Format Journal Article
LanguageEnglish
Published New York Springer US 01.12.2018
Springer Nature B.V
Subjects
Artificial Intelligence
Benchmarks
Clustering
Computer Science
Machines
Manufacturing
Mechanical Engineering
Processes
Unbalance
k-means
Benchmark
Clustering quality
Clustering algorithms
Online AccessGet full text

Cover

More Information
Summary:This paper has two contributions. First, we introduce a clustering basic benchmark. Second, we study the performance of k-means using this benchmark. Specifically, we measure how the performance depends on four factors: (1) overlap of clusters, (2) number of clusters, (3) dimensionality, and (4) unbalance of cluster sizes. The results show that overlap is critical, and that k-means starts to work effectively when the overlap reaches 4% level.
Bibliography:ObjectType-Article-1
SourceType-Scholarly Journals-1
ObjectType-Feature-2
content type line 14
ISSN:0924-669X
1573-7497
DOI:10.1007/s10489-018-1238-7