Table 1.

Discriminant analysis methods for wasting as function of weight-for-age

Discriminant Method	TP*	FP	FN	TN	Se	Sp	Test Error rate
Linear Discriminant Analysis (LDA)	13	8	45	1181	0.224	0.993	0.043
Quadratic Discriminant Analysis (QDA)	13	9	45	1180	0.224	0.992	0.043
K-Nearest Neighbour (KNN)	17	16	41	1173	0.293	0.987	0.046
Epanechnikov Kernel Discriminant (EKD)	13	8	45	1181	0.224	0.993	0.043
Normal Kernel Discriminant (NKD)	9	2	49	1187	0.155	0.998	0.041
Logistic Discriminant Analysis (LDA)	13	9	45	1180	0.224	0.992	0.043
Support Vector Machine (SVM)	9	3	49	1186	0.155	0.997	0.042
Neural networks (NNET)	11	8	47	1181	0.190	0.993	0.040
Mixture Discriminant Analysis (MDA)	15	13	46	1176	0.259	0.989	0.045
Flexible Discriminant Analysis (FDA)	13	8	45	1181	0.224	0.993	0.043
Multivariate Adaptive Regression Splines (MARS)	14	12	44	1177	0.241	0.990	0.045
Adaptive Backfitting (BRUTO)	13	9	45	1180	0.224	0.992	0.043

^*True Positive (TP), False Positive (FP), False Negative (FN), True Negative (TN), Sensitivity (Se), Specificity (Sp)