Fig. 1From: Don’t dismiss logistic regression: the case for sensible extraction of interactions in the era of machine learningCharacteristics of generalized linear modeling approach (blue), decision trees (orange), random forest (green), true model (black); (a-b) Predictor X1 versus binary response Y for the a) linear continuous predictor, b) non-linear transformed predictor; c) binary prediction performance as a function of interaction strength, smoothed using Savitzky-Golay filter to best illustrate trendsBack to article page