Improving biomarker list stability by integration of biological knowledge in the learning process

BMC Bioinformatics

Table 4 Canberra distance and accuracy across breast cancer datasets

	k	GSE2990 vs GSE3494	GSE3494 vs GSE7390	GSE2990 vs GSE7390	Mean Canberra Distance	Mean Accuracy
No prior	4182	0.95	0.94	0.94	0.94	95%
GO BP	4268	0.63	0.65	0.65	0.65	95%
GO MF	3456	0.62	0.62	0.63	0.63	94%
PPI NG	8684	0.62	0.61	0.62	0.62	96%
PPI JA	22207	0.96	0.96	0.97	0.97	95%
PPI FS	22207	0.96	0.97	0.97	0.97	96%
PPI SC	22207	0.91	0.92	0.93	0.93	95%
PE	128	0.70	0.72	0.74	0.74	96%
SP	163	0.68	0.71	0.62	0.62	95%
MI	310	0.62	0.65	0.64	0.64	93%

Pair-wise Canberra distance between the three breast cancer datasets at different number of features selected according to the minimum Canberra distance within datasets, using the standard classification approach without prior knowledge integration and different prior knowledge based similarity matrices. The corresponding mean value and the mean accuracy obtained across the three datasets are also reported.

ISSN: 1471-2105