Spatial prediction of malaria prevalence in Papua New Guinea: a comparison of Bayesian decision network and multivariate regression modelling approaches for improved accuracy in prevalence prediction

Table 3 Area under the curve values for GLM and BDN models

	AUC values
	P. falciparum		P. vivax
GLM village level cross-validation AUC values	Training dataset	Test dataset	Training dataset	Test dataset
GLM village level cross-validation AUC values	0.83	0.49	0.87	0.56
BDN village level cross-validation AUC values	Training dataset	Test dataset	Training dataset	Test dataset
BDN village level cross-validation AUC values	0.74	0.75	0.74	0.76

Area under the curve (AUC) results of Receiver Operating Characteristic (ROC) cross validation for generalised linear regression (GLM) models and Bayesian decision network (BDN) models

ISSN: 1475-2875