Fig. 3

The selection of variables in the OA prediction model was analyzed by LASSO regression. (A) Each curve in the distribution of regression coefficients represents the trajectory of a specific characteristic coefficient. (B) LASSO regression cross-validation profile. The red dot indicates the mean square error corresponding to the λ value, with the vertical axis representing the coefficient values. The number at the top denotes the count of nonzero coefficients in the model, while the horizontal axis shows the logarithmic value of the regularization parameter λ