. Author manuscript; available in PMC: 2010 Oct 26.

Published in final edited form as: J Am Stat Assoc. 2007 Jun 1;102(478):560–572. doi: 10.1198/016214507000000112

Table 3.

Posterior Quantiles for Different Models (corresponding to the models with best L-measure) Under the Three Activation Schemes for the Melanoma Dataset

Parameter	First [model 1(d)]		Last [model 1(b)]		HA–Bin (model 2)		Mixture [model 1(d)]
Parameter	Median	(2.5%, 97.5%)	Median	(2.5%, 97.5%)	Median	(2.5%, 97.5%)	Median	(2.5%, 97.5%)
Intercept	−1.835	(−2.242, −1.479)	−.924	(−1.196, −.675)	.089	(−.130, .255)	−1.665	(−2.156, −1.236)
Age	−.019	(−.244, .215)	−.023	(−.186, .144)	.118	(−.033, .283)	−.024	(−.250, .220)
Sex (male = 0)	−.041	(−.568, .441)	.025	(−.385, .396)	−.164	(−.472, .164)	−.014	(−.582, .472)
Performance status	−.283	(−1.151, .472)	−.246	(−1.023, .372)	−.239	(−.732, .308)	−.362	(−1.226, .407)
η (Weibull link)					−.929	(−1.136, −.749)
Cure fraction (exp(− θ))	.342	(.270, .408)	.359	(.297, .419)			.338	(.250, .404)
ρ	1.426	(1.232, 1.622)	1.193	(1.038, 1.353)	1.192	(1.050, 1.356)	1.477	(1.283, 1.670)