Fig. 4

Bottom x-axes represent the value of the parameter Log (λ) of the lasso regression model, top x-axes represent the number of corresponding non-zero coefficients, and the ordinate is Binomial Deviance (binary classification anomaly), which indicates the error of the model. There are two numerical dotted lines in the picture, the line with the lowest error on the left and the line with few features on the right. We should choose the model corresponding to λ with as few variables and errors as possible. The right vertical dashed line is the best value of the logarithm, and the number of features corresponding to this value is 7