RE: Diagonal elements of correlation matrices
Dear Harald,
This is documented behavior, see the nonmem help entry on correlation matrix
ouput:
The matrix (which is symmetric) is given in lower triangular form. In
this example, the 2x2 matrix, OMEGA, was constrained to be diagonal;
the omitted entries above (.........) indicate that omega(2,1) is not
estimated, and consequently has no corresponding row/column in the
correlation matrix of the estimates.
With NONMEM 7.2, the diagonal elements are equal to the square root of |
the diagonal elements of the covariance matrix (standard error); with |
previous versions they were 1.0.
Hope this helps,
Jeroen
J. Elassaiss-Schaap Senior Principal
Scientist Phone: + 31 412 66 9320
MSD | PK, PD and Drug Metabolism | Clinical PK-PD Mail stop KR
4406 | PO Box 20, 5340 BH Oss, NL
Quoted reply history
From: [email protected] [mailto:[email protected]] On
Behalf Of Ihmsen, Harald
Sent: Tuesday, March 04, 2014 9:33
To: [email protected]
Subject: [NMusers] Diagonal elements of correlation matrices
Dear colleagues,
I have recognized that the diagonal elements of the correlation matrix of
estimates are not equal to 1 as one would expect. I have attached the output of
the theophylline test data set, obtained with NM730. It seems that each
diagonal element in the correlation matrix is just the square root of the
corresponding value in the covariance matrix. The off-diagonal elements,
however, are correct. When looking at older NONMEM versions, this occurred
also in NM720 but not in NM712.
Of course, nobody would really pay attention to the diagonal elements of a
correlation matrix, but it looks odd and should be corrected in future
versions.
Best regards
Harald Ihmsen
Dr. Harald Ihmsen
Anästhesiologische Klinik
Universitätsklinikum Erlangen
Krankenhausstr. 12
91054 Erlangen
Tel: ++49 (9131) 8539157
Fax: ++49 (9131) 8539161
e-mail:
[email protected]<mailto:[email protected]>
************************************************************************************************************************
********************
********************
******************** FIRST ORDER
********************
******************** CORRELATION MATRIX OF ESTIMATE
********************
********************
********************
************************************************************************************************************************
TH 1 TH 2 TH 3 OM11 OM12 OM13 OM22
OM23 OM33 SG11
TH 1
+ 7.08E-01
TH 2
+ 2.27E-01 7.26E-03
TH 3
+ 1.42E-02 9.47E-01 4.53E-03
OM11
+ 9.69E-01 1.78E-01 -1.76E-02 4.83E+00
OM12
+ 4.36E-01 7.12E-01 5.95E-01 3.19E-01 1.40E-02
OM13
+ 9.01E-03 4.60E-01 4.78E-01 -1.38E-01 7.57E-01 4.75E-01
OM22
+ 8.48E-02 4.72E-01 3.86E-01 -1.22E-02 8.22E-01 7.05E-01 1.22E-04
OM23
+ 7.55E-03 -5.66E-02 -4.93E-02 -1.30E-01 4.39E-01 6.65E-01 6.88E-01
3.73E-03
OM33
+ 4.10E-02 -2.42E-01 -1.72E-01 -8.42E-02 1.96E-01 5.11E-01 3.29E-01
8.95E-01 2.12E-01
SG11
+ -1.71E-01 2.61E-01 1.27E-01 -2.68E-01 4.54E-01 3.51E-01 5.85E-01
2.41E-01 -6.81E-02 1.05E-01
Notice: This e-mail message, together with any attachments, contains
information of Merck & Co., Inc. (One Merck Drive, Whitehouse Station,
New Jersey, USA 08889), and/or its affiliates Direct contact information
for affiliates is available at
http://www.merck.com/contact/contacts.html) that may be confidential,
proprietary copyrighted and/or legally privileged. It is intended solely
for the use of the individual or entity named on this message. If you are
not the intended recipient, and have received this message in error,
please notify us immediately by reply e-mail and then delete it from
your system.