Table 4.

AD/NC classification performance for all the frameworks in comparison. Total of 412 samples used. The best value for each performance measure is highlighted in bold. (Acc: Accuracy, Sen: Sensitivity, Spe: Specificity, AUC: area under ROC curve, time: the average imputation time for each fold, rEM: Regularized EM, mFPC: modified FPC)

Framework	Imputation	Acc.	Sen.	Spe.	AUC	Time(s)
Baseline 1	Zero	0.802	0.802	0.804	0.887	0.00
	KNN	0.830	0.801	0.859	0.904	1.76
	rEM	0.816	0.792	0.840	0.892	84.36
	FPC	0.821	0.812	0.831	0.900	115.35
Baseline 2	Zero	0.853	0.843	0.864	0.922	0.00
	KNN	0.868	0.836	0.894	0.927	0.29
	rEM	0.857	0.833	0.879	0.922	23.76
	FPC	0.858	0.843	0.872	0.923	15.31
Proposed	KNN	0.850	0.745	0.936	0.914	0.32
	rEM	0.885	0.837	0.927	0.944	24.20
	mFPC	0.880	0.852	0.904	0.947	0.39