Fig. 2From: Exploring the variable importance in random forests under correlations: a general concept applied to donor organ quality in post-transplant survivalThe figure shows the VIMPs of the 10 variables with highest importance in the Random Survival Forest analysis of UNOS data under model description \(\text {model}_\text {A}\) as well as the corresponding VIMPS under model description \(\text {model}_\text {B}\). Thereby, in \(\text {model}_\text {B}\) KDPI was replaced by its residuals whereas all other variables remained unchanged. Thus, \(\text {VIMP}_\text {B}(\text {KDPI})\) refers to residual information and all other VIMPs to original variables’ information within the two modelsBack to article page