Skip to main content

Table 5 Median values (and associated standard deviation) for the typical bicluster dimension parameters in both data collections: number of regions in each bicluster, number of time points and bicluster area. When comparing this results to the virtual error values, a apparent relation comes between the bicluster size and the associated virtual error, which make sense

From: Biclustering fMRI time series: a comparative study

Algorithms Artificial data Real data
Time points Region points Area Time points Region points Area
Biclustering
Bimax \(6 \pm 7\) \(4 \pm 2\) \(24 \pm 25\) \(6 \pm 2\) \(21 \pm 21\) \(126 \pm 386\)
BicPAM \(7 \pm 3\) \(2 \pm 1\) \(16 \pm 11\) \(27 \pm 8\) \(2 \pm 0\) \(54 \pm 16\)
CCC \(3 \pm 4\) \(3 \pm 2\) \(10 \pm 9\) \(5 \pm 2\) \(4 \pm 9\) \(18 \pm 22\)
FABIA \(43 \pm 5\) \(4 \pm 2\) \(184 \pm 76\) \(29 \pm 4\) \(14 \pm 52\) \(403 \pm 1204\)
Spectral Biclustering \(23 \pm 11\) \(7 \pm 3\) \(136 \pm 112\) \(20 \pm 7\) \(17 \pm 6\) \(330 \pm 211\)
ISA \(14 \pm 9\) \(3 \pm 1\) \(44 \pm 46\) \(7 \pm 5\) \(30 \pm 28\) \(152 \pm 276\)
XMotifs \(49 \pm 13\) \(2 \pm 1\) \(114 \pm 45\) \(19 \pm 11\) \(4 \pm 15\) \(78 \pm 71\)
Region clustering
kmeans 150 \(3 \pm 2\) \(450 \pm 283\) 94 \(4 \pm 1\) \(376 \pm 123\)
pectral 150 \(4 \pm 4\) \(525 \pm 610\) 94 \(3 \pm 117\) \(282 \pm 11038\)
ward 150 \(6 \pm 5\) \(900 \pm 700\) 94 \(3 \pm 1\) \(282 \pm 100\)
Temporal clustering
kmeans \(3 \pm 5\) 26 \(78 \pm 119\) \(3 \pm 0\) 463 \(1389 \pm 197\)
spectral \(3 \pm 6\) 26 \(78 \pm 158\) \(86 \pm 36\) 463 \(39818 \pm 16621\)
ward \(3 \pm 9\) 26 \(78 \pm 247\) \(3 \pm 0\) 463 \(1389\pm 222\)