BMC Public Health

Table 1 Performance metrics of random forest models

From: Predicting self-perceived general health status using machine learning: an external exposome study

	2012 dataset	2016 dataset
N	199,840	244,557
N poor SPGH status	7,081	10,031
N variables	81	91
MTry	9	9
N trees	1000	1000
AUC^a (95% CI)	0.864 (0.852– 0.876)	0.890 (0.885—0.895)
Sensitivity^b (95% CI)	0.777 (0.753 – 0.800)	0.811 (0.789 – 0.833)
Specificity^c (95% CI)	0.818 (0.793 – 0.842)	0.837 (0.820 – 0.854)

^aAverage AUC of five-fold crossed validated random forest models
^bAverage of five-fold crossed validated random forest models
^cAverage specificity of five-fold crossed validated random forest models

Back to article page

ISSN: 1471-2458

Contact us

Submission enquiries: bmcpublichealth@biomedcentral.com
General enquiries: ORSupport@springernature.com