. Author manuscript; available in PMC: 2020 Jun 1.

Published in final edited form as: J Biomed Inform. 2019 Apr 25;94:103185. doi: 10.1016/j.jbi.2019.103185

Table 2.

Performance metrics of random forest model using OMOP + NLP datasets with binary features.

severity	AUC (std. dev.)	PPV (precision)	Sensitivity (recall)	Accuracy	MAE
false positive	0.83 (0.08)	0.527	0.557	0.779	0.729
mild	0.841 (0.059)	0.56	0.431	0.802	0.569
moderate	0.9 (0.064)	0.727	0.821	0.809	0.291
severe	0.982 (0.017)	0.881	0.804	0.953	0.239
all	0.893 (0.031)	0.674	0.653	0.836	0.446