Figure 1From: A random forest approach to the detection of epistatic interactions in case-control studiesPrinciples of epi Forest. In the first stage, a random forest is trained with all SNPs to obtain the gini importance of each SNP, and a sliding window sequential forward feature selection (SWSFS) algorithm is used to select a subset of candidate SNPs that can minimize the classification error. In the second stage, statistical tests on the basis of the B statistics are applied to detect significant one-, two-, and three-way epistatic interactions.Back to article page