Figure 2
From: Identification of multiple gene-gene interactions for ordinal phenotypes
Comparison between OMDR and MDR. Performance of OMDR and MDR is compared via EP (empirical power), average GCVC (K = 1), average TSTB (testing tau-b), average TRTB (training tau-b), and their average values over models. Note that true causal factor was a two-locus classifier (i.e., two-way interaction), and all two-locus classifiers were searched by both methods.