Skip to main content

Table 5 Comparison of Cox regression and RSF models in different scenarios

From: Risk factors associated with major adverse cardiac and cerebrovascular events following percutaneous coronary intervention: a 10-year follow-up comparing random survival forest and Cox proportional-hazards model

Included variable Model C index OOB error rate IBS*
All 13 variables Cox 0.626 0.374 0.135
RSF 0.648 0.352 0.124
4 variables with significant unadjusted hazard ratios (diabetes, smoking, age, stent length) Cox 0.670 0.328 0.126
6 variables with positive VIMP (diabetes, smoking, age, stent length, setting, hyperlipidemia RSF 0.697 0.317 0.112
The same variables (diabetes, smoking, age, stent length) Cox 0.670 0.328 0.126
RSF 0.684 0.312 0.113
  1. Bold values are the higher values of indices for comparing two models
  2. *Integrated brier score