Re: [NMusers] Estimation method for ETAs with POSTHOC

From: Rong Chen <>
Date: Sun, 7 May 2017 10:55:24 +0000 (UTC)

Hi Tingjie,I think this article could answer your question by indicating th=
e objective function used to estimate post hoc eta and how to reproduce the=
 estimate process in R. There were R codes example of a one-compartment PK =
model with 3 etas to be estimated. Title: R-based reproduction of the =
estimation process hidden behind NONMEM® Part 1: first-order approxima=
tion. URL:
Best wishes,Rong

      From: "Guo, Tingjie" <>
 To: nmusers <>
 Sent: Friday, 5 May 2017, 20:17
 Subject: [NMusers] Estimation method for ETAs with POSTHOC
Dear NMusers,

I have two questions about the mathematical details on POSTHOC estimation i=
n NONMEM. How does NONMEM actually do when doing POSTHOC ($EST METHOD=0 M=
AXEVEL=0) to get ETAs? I assume but not for sure there is an objective fu=
nction to be minimized, somewhat like:
      Objective function = SUM(Yobs-Ypred)^2/sigma^2 + SUM=

Is there an objective function used in NONMEM when doing POSTHOC? If so, wh=
at is that function?

Secondly, if the answer to above is yes, let’s assume the real obje=
ctive function is the same as what I mentioned above. I wonder how (NONMEM =
does) to minimize this function? I am currently trying to do similar things=
 in R/Python language. I tried Metropolis-Hasting algorithm and Simulated A=
nnealing algorithm, but with some technical problems. And the result was no=
t comparable with NONMEM as well. Can someone give me a direction for this?=
 The more mathematically detailed the better.


Warm regards,


Received on Sun May 07 2017 - 06:55:24 EDT

This archive was generated by hypermail 2.3.0 : Fri Sep 27 2019 - 16:54:58 EDT