Bioequivalence and Bioavailability Forum • Speeed!

Speeed! [Two-Stage / GS Designs]

posted by Helmut – Vienna, Austria, 2015-02-20 17:58 (4130 d 00:57 ago) – Posting: # 14476
Views: 14,763

Hi nobody,

❝ Yeeaaaaahhhh, now I remember that... even ran the code in R sometimes... ;-)

Wouldn’t recommend that any more. If you want I can send you my current version (including futility criteria). Parallel designs not implemented yet.

In the primary step instead of assessing the maxima for each n₁, I first look at the global maximum and run the second step for each TIE exceeding f.i. 95% of the global one. That’s better especially for “type 2” designs which have a flat TIE surface.
Anders’s linear model is not always the best. Now I fit both a linear and quadratic and select the better one based on the minimum AIC.

❝ ❝ ...With a SETI-like approach we could even think about TSDs for RSABE…

❝

❝ Let's start with distributed computing in, let's say 10 to 20 instances on different computers..

Once the API is ready… On my machine I often use three instances of R without problems.

—
Dif-tor heh smusma 🖖🏼 Довге життя Україна!
Helmut Schütz

The quality of responses received is directly proportional to the quality of the question asked. 🚮
Science Quotes