Skip to main content
Fig. 2 | Journal of Translational Medicine

Fig. 2

From: Prediction of acute kidney injury after cardiac surgery: model development using a Chinese electronic health record dataset

Fig. 2

Feature selection by the sliding windows sequential forward feature selection method. First, variable importance score (VIS) of features was obtained from a random forest algorithm and ranked in descending order. The variables were then included one by one to the random forest model in order of VIS rank. Finally, the optimal number of variables (14 variables) was determined by minimum model error (red circle)

Back to article page