Table 3.

Time-varying selection results of separate-penalty-varying-coefficient models (Method 2) for variables 2, 3, and 8 from 100 runs.

n	cp	10 covariates						20 covariates
		X2		X3		X8		X2		X3		X8
		Int	TV	Int	TV	Int	TV	Int	TV	Int	TV	Int	TV
s = 1
200	20	63	12	97	67	100	1	56	11	96	64	100	0
	40	81	26	86	76	98	3	80	36	87	71	100	4
400	20	94	43	94	89	99	1	98	56	93	91	100	1
	40	97	71	89	82	100	1	100	84	90	80	100	4
s = 2
200	20	98	79	85	98	100	1	98	78	75	100	100	0
	40	100	95	83	97	100	4	100	97	79	99	100	3
400	20	100	100	85	100	100	2	100	100	90	100	100	3
	40	100	100	86	100	100	3	100	100	88	100	100	10

Int: the intercept component. TV: the time-varying component.