Table 2.
Silhouette index values for clustering solutions produced by PAM with different values of K
| Data set | K | Silho Eucli-PAM | Silho City-PAM | Silho Cosine-PAM |
|---|---|---|---|---|
| Yeast | 5 | 0.3792 | 0.367 | 0.381 |
| 10 | 0.4531 | 0.452 | 0.442 | |
| 20 | 0.4415 | 0.437 | 0.435 | |
| 30 | 0.4075 | 0.411 | 0.426 | |
| 40 | 0.40 | 0.421 | 0.423 | |
| 50 | 0.397 | 0.432 | 0.419 | |
| Multiple tissues | 5 | 0.354 | 0.361 | 0.359 |
| 10 | 0.383 | 0.372 | 0.368 | |
| 20 | 0.394 | 0.379 | 0.382 | |
| 30 | 0.406 | 0.394 | 0.392 | |
| 40 | 0.4299 | 0.419 | 0.404 | |
| 50 | 0.429 | 0.402 | 0.418 | |
| 60 | 0.415 | 0.398 | 0.416 | |
| 70 | 0.414 | 0.391 | 0.409 |
The data in boldface represents optimal value of ‘K’ i.e. dimension of gene space corresponding to optimal Silhouette index for all of three distance based PAM versions