Fig. 4From: Sparsifying machine learning models identify stable subsets of predictive features for behavioral detection of autismHistogram of predicted probabilities for logistic regression with L 2 penalization (a) and linear SVM with L 1 penalization (b) on the test set. The x-axis denotes the predicted probabilities for logistic regression and the decision function for SVM and the y-axis represents the number of subjects. Each subplot correspond to a different labelBack to article page