Fig. 2From: Improving machine learning reproducibility in genetic association studies with proportional instance cross validation (PICV)Consistency of training and testing performance measures for models with and without the interaction term, comparing a traditional cross validation procedure to PICV. Experimental scenario in which both SNPs have a MAF of 0.5, n = 2000. PPV: positive predictive value, NPV: negative predictive value, N.S.: not significantBack to article page