Figure 2From: Scientific knowledge is possible with small-sample classificationRMS and maxBayes(λ) . (a) RMS (y-axis) as a function of the Bayes error (x-axis) for leave-one-out with dimension D=10 and sample sizes n=20 (plus sign), 40 (triangle), 60 (circle); (b) maxBayes(λ) curves corresponding to the RMS curves in part (a).Back to article page