Skip to main content

Table 4 Test results for real expression data sets.

From: MULTI-K: accurate classification of microarray subtypes using ensemble k-means clustering

Data set

MULTI-K

GCCk

GCCc

k-means

Hier.

    

gap

Sil

gap

Sil

Leukemia (3)

0.7364 (2)

0.7284 (4)

0.7040 (2)

0.4604 (6)

0.6930 (2)

0.7521 (4)

0.4881 (2)

Lymphoma (3)

1.0000 (3)

0.3016 (5)

0.9414 (2)

0.1454 (6)

0.8943 (2)

0.9414 (2)

0.9414 (2)

Colon (2)

1.0000 (2)

0.0349 (5)

0.0386 (4)

0.0540 (6)

0.0088 (25)

0.8944 (2)

0.8944 (2)

Thyroid I (2)

1.0000 (2)

0.2396 (5)

0.1351 (4)

0.7335 (3)

0.7772 (2)

0.3462 (3)

0.4183 (2)

Thyroid II (2)

1.0000 (2)

0.4468 (4)

0.3986 (5)

1.0000 (2)

1.0000 (2)

0.5895 (4)

1.0000 (2)

St. Jude (6)

0.8015 (5)

0.7490 (5)

0.6697 (4)

0.5579 (9)

0.1985 (2)

0.1985 (2)

0.1985 (2)

Normal I (12)

1.0000 (12)

1.0000 (12)

0.4964 (11)

0.9546 (13)

0.7356 (10)

0 (1)

1.0000 (12)

Normal II (13)

0.7977 (10)

0.7629 (11)

0.2476 (13)

0.7650 (15)

0.6875 (10)

0.1985 (2)

0.6830 (16)

  1. In each parenthesis is shown the known (left header) and the predicted number of clusters. ARI values that exceed 0.7 are shown in bold.