Table 2.

Analysis of the PCR data. Marginal and Joint refer to the variable selection procedures (Bondell and Reich, 2012) based on Bayesian marginal credible set and Bayesian joint credible set, respectively. MS is the average size of the selected model. FS is the number of frequently selected variables, i.e., that were selected at least 95 times in 100 repetitions. TS refers to the total number of variables selected at least once from 100 repetitions. Standard errors are provided in parenthesis.

Method	MSPE	MS	FS	TS
piMoM(MAP)	0.283 (0.17)	1.00 (0.00)	1	1
piMoM(LS)	0.282 (0.17)	1.00 (0.00)	1	1
peMoM(MAP)	0.291 (0.18)	1.02 (0.14)	1	2
peMoM(LS)	0.287 (0.17)	1.02 (0.14)	1	2
g-prior	0.368 (0.20)	4.07 (0.56)	1	133
lasso	0.542 (0.39)	17.97 (8.62)	1	211
scad	0.308 (0.23)	12.66 (7.62)	2	163
mcp	0.308 (0.21)	2.20 (0.94)	0	29
Marginal(p = 2000)	0.456 (0.40)	17.47 (11.16)	0	273
Joint(p = 2000)	0.440 (0.40)	16.42 (11.06)	1	185