Figure 5.

Mean accuracy of weigthed voting classifier versus mean BSS/WSS. Mean accuracy of the weighted voting classifier using three features versus the mean BSS/WSS of these features for two gene expression data sets (leukemia, lung cancer) and a proteomics data set (prostate cancer). Mean values were calculated across 1,000 random training:test set partitions. Features to include in the classifiers were identified through a jackknife procedure through which features were ranked according to their frequency of occurrence in the top 1% most significant features based on t-statistics across all jackknife samples. Mean BSS/WSS was calculated separately using the training and test set portions of each random partition.