Fig. 2

Correlation of feature variables and feature selection process in machine learning survival prediction models. (A) Heat map of the Spearman correlation matrix of the characteristic variables. The color shade indicates the strength of the correlation. The size of the circle indicates the strength of the correlation. The asterisk indicates the significance level of the correlation. (B) LASSO regression paths show the coefficients of the variables at different values of the regularization parameter (λ). (C) Cross-validation error map for selecting the optimal λ in LASSO. The vertical dashed line indicates the optimal λ for realizing the cross-validation error. (D) Importance of variables based on Boruta’s algorithm, where attributes are categorized as “confirmed” (red) and “rejected” (brown). (E) Venn diagrams compare the variables selected by the three different methods, showing the overlap of the selected variables.