Machine learning approach for predicting under-five mortality determinants in Ethiopia: evidence from the 2016 Ethiopian Demographic and Health Survey

Genus

Journal of Population Sciences

Table 2 Model accuracy metrics for all models as evaluated on the test data

Confusion matrix		Random Forest		Logistic regression		KNN model
		Predicted		Predicted		Predicted
		Alive	Dead	Alive	Dead	Alive	Dead
Observed	Alive	698	343	632	418	475	566
	Dead	12	28	16	24	14	26
			%		%		%
Accuracy			67.2		59.9		46.3
Sensitivity			98.3		97.5		97.1
Specificity			7.5		5.34		4.4
Positive predictive value			70.0		59.9		45.6
Negative predictive value			67.0		60.0		65.0
AUC			72.0		66.1		55.5