BMC Medical Informatics and Decision Making

Table 5 Gradient boosting machine error metrics for patient 30-day mortality model on frequent subgraphs representing patient subgroups

From: Unsupervised phenotyping of sepsis using nonnegative matrix factorization of temporal trends from a multivariate panel of physiological measurements

Measure	Train	Test
Accuracy	0.892	0.863
Precision	0.885	0.802
Recall	0.726	0.681
F-score	0.773	0.716
AUC	0.726	0.681

Back to article page

ISSN: 1472-6947

Contact us

General enquiries: journalsubmissions@springernature.com