keep it simple! [🇷 for BE/BA]

posted by ElMaestro  – Denmark, 2015-04-20 17:39 (4076 d 09:49 ago) – Posting: # 14716
Views: 31,848

Hi Helmut,

❝ Statler & Waldorf-day today?


Every day is a Statler & Waldorf day :-D Not good for my blood pressure. Today's major trouble is bacterial DNA, though. Why do people publish sensational papers about having sequenced a new bacterial species when they are not publishing the genome? :confused:

❝ cnames <- c("subj", "drug", "seq", "prd", "Cmax", MeanR),

❝ (...) "\nT:", sprintf("%.6f", MeanT), "\n")


Ohdearohgoodnessmelordhavemercy. Yuck. :ponder:

❝ I’m not promoting SAS. :-D


Don't worry. I know that for a long time already.

There is still something that really bothers me and that is the log Likelihood difference of R versus WNL. With REML I think optimisation switches back and forth be in terms of the estimates. You fiddle with the covariance matrix and read the logLikelihood. Then you fiddle with the fixed effects and read the logLikelihood. Then you fiddle with the covariance matrix, and so forth. Repeat until some convergence criterion is met. ML would only be plain and simple covariance matrix fiddling, I think (?). Thus I have a vague and completely unsubstantiated feeling this could possibly be fixed by proper inputs and that you'd then see agreement about the likelihoods and probably the treatment effects too. It is a guess, and undertunately I do not know my way around with R or mixed models to an extent where I can do it.
—
Pass or fail!
ElMaestro
Thread locked

Complete thread:

UA Flag
Activity
 Admin contact
23,654 posts in 4,992 threads, 1,571 registered users;
186 visitors (0 registered, 186 guests [including 16 identified bots]).
Forum time: 03:28 CEST (Europe/Vienna)

“Data! Data! Data!” he cried impatiently.
“I can’t make bricks without clay!”    Arthur Conan Doyle

The Bioequivalence and Bioavailability Forum is hosted by
BEBAC Ing. Helmut Schütz
HTML5