Dear All,
reading an article by Rodolphe Thiebaut and Helene Jacqmin-Gadda ("Mixed
models for longitudinal
left-censored repeated measures") I have found this program in SAS
proc nlmixed data=TEST QTOL=1E-6;
parms sigsq1=0.44 ro=0.09 sigsq2=0.07 sigsqe=0.18 alpha=3.08 beta=0.43;
bounds $B!](B1< ro < 1, sigsq1 sigsq2 sigsqe >= 0;
pi=2*arsin(1);
mu=alpha+beta*TIME+a i+b i*TIME;
if OBS=1 then ll=(1/(sqrt(2*pi*sigsqe)))*exp(-(RESPONSE-mu)**2/(2*sigsqe));
if OBS=0 then ll=probnorm((RESPONSE-mu)/sqrt(sigsqe));
L=log(ll);
.......
I haven't found a simple solution for the conditional LL:
if OBS=1 then ll=(1/(sqrt(2*pi*sigsqe)))*exp(-(RESPONSE-mu)**2/(2*sigsqe));
if OBS=0 then ll=probnorm((RESPONSE-mu)/sqrt(sigsqe));
using nlme.
Any suggestion will be appreciated.
Giovanni
sessionInfo()
R version 2.6.2 (2008-02-08)
i386-pc-mingw32
locale:
LC_COLLATE=Italian_Italy.1252;LC_CTYPE=Italian_Italy.1252;LC_MONETARY=Italian_Italy.1252;LC_NUMERIC=C;LC_TIME=Italian_Italy.1252
attached base packages:
[1] splines tools stats graphics grDevices utils datasets
methods base
other attached packages:
[1] JM_0.1-0 MASS_7.2-41 Design_2.1-1 survival_2.34-1
nlme_3.1-88
[6] Hmisc_3.4-3
loaded via a namespace (and not attached):
[1] cluster_1.11.10 grid_2.6.2 lattice_0.17-6
[[alternative HTML version deleted]]
