Figure 4From: A unifying theory for genetic epidemiological analysis of binary disease dataROC curve for predicting disease status using an IGE model. Data from simulation of the SIR model with variation in susceptibility and infectivity; curves in green = the probability function with lambda estimated as in equation (15) used as classification criterion (AUC = 0.964); brown (overlapping with green curve) = the probability function with lambda estimated using the Taylor expansion from equation (25) used as classification criterion (AUC = 0.964); purple = an IGE model (equation (26)) used as classification criterion (AUC = 0.751); black = random classification; grey = perfect classification.Back to article page