Displaying 2 results from an estimated 2 matches for "gcomp".
Did you mean:
comp
2013 Mar 10
0
Steepest Ascent Algorithm
...ikelihood function
#w = 0 or 1 depending on whether they were censored or not
#t is the suvival time
#d is 0 if they were the control group and 1 if they were in the treatment
group
#theta is a vector of paramters (a,Bo, and B1 over which to maximize upon
lordata[4,3]
like<-function(theta,data,gcomp=FALSE,hesscomp=FALSE)
{
a = theta[1]
Bo=theta[2]
B1=theta[3]
d=data[,1] #treatment
w=data[,2] #censored
t=data[,3] #survival time
mu = (t^a)*exp(Bo+d*B1)
l = sum(w*log(mu)-mu+w*log(a/t))
if(gcomp==TRUE) {
grad=matrix(0,3,1)
grad[1]=sum(w*log(t)-mu*log(t...
2006 Sep 01
0
lmer applied to a wellknown (?) example: summary
...or for pointing me to the thread
'Doubt about nested aov output' where the rat-example was hiding...:
http://search.gmane.org/?query=Doubt+about+nested+aov+output&email=&group=gmane.comp.lang.r.general&sort=revdate&DEFAULTOP=and&xP=doubt.nested.aov.output.&xFILTERS=Gcomp.lang.r.general---A
In this great thread you find a description not only on using aov on
this nested rat-data, but also how to handle it with lmer, e.g. model
specification and coding of grouping levels.
Best regards,
Henrik
--
************************
Henrik P?rn
Department of Biology
NTNU...