Figure 6From: A random forest approach to the detection of epistatic interactions in case-control studiesClassification error rates of random forests using up to the first 2,000 most important SNPs. x-axis is the number of SNP markers used. y-axis is the classification error rate given by the random forest (one million trees). The circle (with 84 markers and an error rate of 8.5%) represents the minimum in this curve.Back to article page