Table 5. Performance of the considered classification methods on three gene expression microarray datasets.

	ridge.net	ARACNE	KEGG
Colon	67.857 (11.77)	70.357 (11.37)	66.143 (12.17)
Lung	59.413 (5.88)	56.457 (6.31)	56.37 (5.83)
Prostate	87.441 (6.09)	87.029 (5.40)	84.353 (6.78)

The graphs integrated in the classification methods NB-SVM and gCDA were either inferred with two methods, ridge.net and ARACNE, or extracted from KEGG. In this table are presented the mean (standard deviation) of the good classification rate over 100 MCCV iterations.