Table 20 Rank of similarity measures based on clustering results
From: A set theory based similarity measure for text clustering and classification
Measure/table | Purity | Completeness | Rand index | -Calinski-Harabasz Index | Davies-Bouldin Index | Point Total out of 20 | Rank |
|---|---|---|---|---|---|---|---|
Euclidean | 4 | 2 | 2 | 0 | 2 | 10 | 2 |
Cosine | 3 | 2 | 3 | 1 | 0 | 9 | 4 |
Jaccard | 0 | 0 | 0 | 4 | 0 | 4 | 7 |
Bhattacharya | 1 | 3 | 1 | 0 | 1 | 6 | 6 |
kullback–Leibler | 0 | 0 | 0 | 4 | 4 | 8 | 5 |
Manhattan | 0 | 0 | 0 | 0 | 4 | 4 | 7 |
PDSM | 2 | 2 | 2 | 2 | 1 | 9 | 3 |
STB-SM | 2 | 4 | 4 | 1 | 0 | 11 | 1 |