Table 3.

Evaluation of classification algorithms on AD data.

Classification algorithm	Feature selection	Accuracy	Sensitivity	Specificity	ROC-AUC	prAUC
LR	All genes	0.737	0.811	0.635	0.821	0.848
	Knowledge genes	0.713	0.783	0.615	0.802	0.830
	VSSRFE	0.733	0.755	0.702	0.812	0.842
	LASSO	0.777	0.790	0.760	0.859	0.899
	VAE	0.648	0.657	0.635	0.661	0.692
SVM	All genes	0.769	0.853	0.654	0.842	0.860
	Knowledge genes	0.737	0.797	0.654	0.800	0.822
	VSSRFE	0.745	0.769	0.712	0.827	0.858
	LASSO	0.769	0.797	0.731	0.858	0.898
	VAE	0.579	0.497	0.692	0.615	0.661
XGBoost	All genes	0.599	0.599	0.654	0.724	0.764
	Knowledge genes	0.741	0.713	0.779	0.841	0.875
	VSSRFE	0.794	0.853	0.712	0.847	0.883
	LASSO	0.725	0.587	0.913	0.858	0.902
	VAE	0.628	0.839	0.337	0.660	0.709
RF	All genes	0.741	0.748	0.731	0.820	0.855
	Knowledge genes	0.700	0.720	0.673	0.792	0.820
	VSSRFE	0.810	0.818	0.798	0.889	0.919
	LASSO	0.717	0.573	0.913	0.860	0.903
	VAE	0.656	0.790	0.471	0.678	0.684
MLP	All genes	0.761	0.839	0.654	0.838	0.873
	Knowledge genes	0.721	0.790	0.625	0.803	0.829
	VSSRFE	0.757	0.804	0.692	0.828	0.863
	LASSO	0.765	0.720	0.827	0.855	0.890
	VAE	0.514	0.378	0.702	0.567	0.659
CNN		0.765	0.895	0.587	0.810	0.845
VAE		0.757	0.923	0.529	0.798	0.816