Skip to main content


Figure 2 | BMC Bioinformatics

Figure 2

From: Combining multiple hypothesis testing and affinity propagation clustering leads to accurate, robust and sample size independent classification on gene expression data

Figure 2

The overall performance of the FS methods according to the AUC metric. We have sorted the methods, except the Rnd, which is not actually a method, according to the mean of the AUC values. The standard deviation across all diseases quantifies the robustness of each method. The mean value per disease across all feature selection methods is a difficulty index of discrimination. The NM from the myopathies and the prostate cancer were the most difficult cases towards the phenotype discrimination.

Back to article page