Figure 3From: Importance of data structure in comparing two dimension reduction methods for classification of microarray gene expression dataProstate dataset. Projection of the 102 patients from the prostate dataset (50 without and 52 with tumor) on the discriminant axis obtained with BGA (x-axis), along their coordinates on the first (on the top) and the second (on the bottom) within-group PCA component (y-axis), respectively. For a better legibility, the groups were labeled 0 (for non-tumor prostate samples) and 1 (for tumor prostate samples). Only the 50 most differential genes among 12625 were used for these graphs.Back to article page