Figure 2From: Optimal classifier selection and negative bias in error rate estimation: an empirical study on high-dimensional prediction Subsample analyses. Alon's colon cancer data (left) and Singh's prostate cancer data (right). Boxplots of the minimal error rate over the 124 classifiers for each subsample size (each boxplot corresponds to 20 error rate estimates). Main conclusion: The median minimal error rate does not seem to increase with decreasing sample size.Back to article page