Table 2. Predictive performances of the limited sampling equations for AUC with various sampling time points.

Sampling point		R2	Model equations for AUC	Prediction error (%)			AIC
				ME	MAE	RMSE
One	C0	0.673	C0x15.704+61.203	-3.169	14.531	18.113	536.5
	C2	0.692	C2x6.626+34.433	-2.410	12.618	17.560	531.6
	C4	0.896	C4x8.838+23.165	-1.070	7.633	10.226	446.2
	C6	0.914	C6x10.364+20.787	-0.881	6.888	9.268	430.7
	C8	0.875	C8x11.225+22.443	-1.110	8.026	11.185	460.4
	C12	0.848	C12x15.053+14.557	-1.002	9.714	12.345	476.0
Four	C0, C2, C4, C6	0.978	C0x2.232+C2x1.998+C4x2.539+C6x4.718+4.890	-0.274	4.064	4.784	329.1
Three	C2, C4, C6	0.973	C2x2.136+C4x2.699+C6x5.397+4.827	-0.277	4.293	5.277	343.6
	C0, C2, C6	0.969	C0x2.494+C2x2.292+C6x7.051+5.642	-0.334	4.563	5.691	355.5
	C0, C2, C4	0.954	C0x3.715+C2x2.127+C4x5.654+8.251	-0.501	5.700	6.849	384.8
	C0, C4, C6	0.950	C0x3.715+C2x2.127+C4x5.654+8.251	-0.501	5.700	7.187	392.4
Two	C2, C6	0.962	C2x2.468+C6x7.981+5.623	-0.342	4.926	6.204	368.2
	C4, C6	0.940	C4x3.981+C6x6.079+16.603	-0.678	5.633	7.820	404.8
	C2, C4	0.939	C2x2.415+C4x6.773+9.014	-0.567	6.649	7.887	406.2
	C0, C6	0.929	C0x3.677+C6x8.743+19.219	-0.813	6.133	8.503	418.0
	C0, C4	0.922	C0x4.747+C4x7.094+20.024	-0.909	6.796	8.885	425.0
	C0, C2	0.834	C0x9.352+C2x4.150+21.921	-1.178	9.801	12.972	484.8

AUC, area under the curve; ME, mean error; MAE, mean absolute error; RMSE, root mean squared error; AIC, Akaike’s information criterion.