Hi all,
I've written some function that is substantially an iterative call to glm()
where the linear predictor changes at each iteration with respect to just
one variable.
I used obj$dev as covergence criterion, i.e.:
epsilon<-abs((dev.old-dev.new)/dev.old)
while(epsilon>toll).....
and I used cat(....) to see the trace of the process:
cat(iter=it,dev=dev.new, labels="iteration", fill=T)
[ Is there any best way to print this information? ]
Everything works but with some dataset (really very little) I noted that
the deviance is not always a decreasing function, in sense that there some
iteration where the dev is smaller than one found when the algorithm attains
convergence.
Which solution have I to assume? Is it right to assume the solution when the
increment in deviance is negligible?
thanks for your help,
vito
-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-
r-help mailing list -- Read http://www.ci.tuwien.ac.at/~hornik/R/R-FAQ.html
Send "info", "help", or "[un]subscribe"
(in the "body", not the subject !) To: r-help-request at
stat.math.ethz.ch
_._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._