Ensemble Merit Merge Feature Selection for Enhanced Multinomial Classification in Alzheimer’s Dementia

Table 3

Results for the 3-class dataset with ensemble of NB, J48, RF, SVM and with CPEMM ensemble.

	NB Ensemble			J48 Ensemble			RF Ensemble			SVM Ensemble			NB-CPEMM			J48-CPEMM			RF-CPEMM
	Acc	Pre	Rec	Acc	Pre	Rec	Acc	Pre	Rec	Acc	Pre	Rec	Acc	Pre	Rec	Acc	Pre	Rec	Acc	Pre	Rec
NSDS	0.838	0.854	0.838	0.911	0.895	0.889	0.936	0.941	0.799	0.647	0.564	0.677	0.899	0.904	0.912	0.932	0.958	0.933	0.936	0.941	0.799
NIDS	0.756	0.737	0.556	0.936	0.941	0.937	0.545	0.645	0.545	0.667	0.685	0.734	0.834	0.895	0.738	0.904	0.934	0.907	0.545	0.645	0.545
CIDS-I	0.863	0.833	0.823	0.916	0.933	0.923	0.963	0.963	0.963	0.673	0.699	0.788	0.896	0.812	0.822	0.987	0.955	0.963	0.963	0.963	0.963
CIDS-II	0.756	0.654	0.563	0.945	0.931	0.924	0.797	0.723	0.634	0.685	0.676	0.657	0.885	0.823	0.813	0.971	0.965	0.923	0.888	0.856	0.833

(NSDS: Neuropsychological data set, NIDS: Neuroimaging dataseta, CIDS: combined dataset).
Acc: Accuracy, Pre: precision, Rec: Recall.