Skip to main content


Figure 3 | BMC Bioinformatics

Figure 3

From: A comprehensive sensitivity analysis of microarray breast cancer classification under feature variability

Figure 3

Impact of perturbation variability on feature selection for the Affymetrix datasets. Each dataset was split 50 times into a training and validation set, for which the validation set was subsequently discarded. Ranking was done only on the training sets. In addition, for each training set 50 perturbed versions were created and for each perturbation the overlap between Fn, m, kand and between and F2n, m, kwas determined, yielding 50·50·2 = 5000 overlap estimates for each list size n. The blue curves provide for each n {1,...,100} the mean overlap taken over all corresponding estimates. The red curves indicate the associated average relative strengths between the feature sets Fn, m, kand .

Back to article page