Skip to main content

Table 3 Impurity-based feature selection using Random Forest for predicting \(PA_{f}\) (a) and \(WAI_{f}\) (b)

From: Exploratory application of machine learning methods on patient reported data in the development of supervised models for predicting outcomes

\(PA_{f}\) \(WAI_{f}\)
Model MAE ± SD \({R}^2\) MR Model MAE ± SD \({R}^2\) MR
LR 1.54 ± 1.18 0.25 0.050 LR 1.16 ± 1.12 0.27 0.003
PAR 1.54 ± 1.19 0.25 − 0.087 PAR 1.10 ± 1.14 0.28 − 0.288
SGDR 1.55 ± 1.17 0.25 0.143 SGDR 1.10 ± 1.13 0.29 − 0.243
RFR 1.57 ± 1.13 0.25 0.199 RFR 1.09 ± 1.20 0.25 − 0.246
ABR 1.60 ± 1.14 0.23 0.0 ABR 1.21 ± 1.20 0.18 − 0.090
SVR 1.53 ± 1.15 0.27 0.102 SVR 1.11 ± 1.15 0.27 − 0.221
XGB 1.55 ± 1.13 0.26 − 0.015 XGB 1.18 ± 1.12 0.25 0.016
  1. The best performing model are highlighted in bold letters