TABLE 2.

Performances of PLSR in the four sets of estimations.

The first set: multi-label learning, five main labels
	Age	Education	CSFC	CSCC	CSOC
R	0.627	0.395	0.369	0.585	0.536
RMSE	2.908	1.636	10.727	8.021	12.184
The second set: multi-label learning, five Supplementary labels
	Grip strength	Reading recognition	Picture vocabulary	VSPLOT	Gender∗
R	0.701	0.522	0.555	0.376	97.6%(ACC)
RMSE	8.066	9.038	7.871	4.109	0.996(AUC)
The third set: multi-label learning, ten labels
	Age	Education	CSFC	CSCC	CSOC
R	0.625	0.400	0.367	0.573	0.528
RMSE	2.914	1.629	10.738	8.101	12.278
	Grip strength	Reading recognition	Picture vocabulary	VSPLOT	Gender∗
R	0.704	0.519	0.546	0.382	97.8%(ACC)
RMSE	8.033	9.059	7.920	4.093	0.996(AUC)
The fourth set: single-label learning, five main labels
	Age	Education	CSFC	CSCC	CSOC
R	0.635	0.402	0.380	0.584	0.525
RMSE	2.886	1.628	10.667	8.033	12.281

^∗For gender classification, the classification accuracy (ACC) and the area under curve (AUC) were provided.