Abstract | ||
---|---|---|
•Internal cluster validity indices are often used to determine grouping quality.•We consider Calinski-Harabasz, Dunn, Davies-Bouldin, Silhouette, and other indices.•A large battery of benchmark data sets is studied.•Many indices promote clusterings that match expert knowledge quite poorly.•This calls the use of such indices in clustering algorithm evaluation into question. |
Year | DOI | Venue |
---|---|---|
2021 | 10.1016/j.ins.2021.10.004 | Information Sciences |
Keywords | DocType | Volume |
Clustering methodology,Cluster validity index,Dunn index,Nearest neighbours (NNs),Ordered weighted averaging (OWA) operator,No free lunch | Journal | 581 |
ISSN | Citations | PageRank |
0020-0255 | 0 | 0.34 |
References | Authors | |
0 | 3 |
Name | Order | Citations | PageRank |
---|---|---|---|
Marek Gagolewski | 1 | 0 | 0.34 |
Maciej Bartoszuk | 2 | 17 | 3.62 |
Anna Cena | 3 | 18 | 4.21 |