Skip to main content
Figure 2 | BMC Bioinformatics

Figure 2

From: Integrative disease classification based on cross-platform microarray data

Figure 2

Classification performance increases with size and phenotype homogeneity of disease classes. The disease classes were divided into bins (a) based on the number of datasets (from 3 to 7) in the classes, or (b) based on the p-value of the phenotype distance score (p-value intervals were chosen as: 1.0E-6 to 1.0E-5, 1.0E-5 to 1.0E-4, 1.0E-4 to 1.0E-3, 1.0E-3 to 1.0E-2, and 1.0E-2 to 5.0E-2). For each bin, the average accuracy was calculated by performing ManiSVM and SVM classification.

Back to article page