Re: Help: Non-positive semi-definite message

From: Nick Holford <n.holford@auckland.ac.nz> Subject: Re: Help: Non-positive semi-definite message Date: Tue, 21 Aug 2001 10:14:15 +1200 Ken, Like Alan I have some questions about your suggestion. I have difficulty accepting the interpretation of the model when you fix the correlation at 1. This means that if CL goes up 50% in patient then V will go up exactly 50% in the same patient. I have no prior expectation that this would ever happen. In the case we are discussing were the data does not allow the covariance between CL and V to be estimated adequately I feel more comfortable with the the assumption that the covariance is 0. A third middle ground possibility would be to include a third ETA to define the covariance as a random but FIXED parameter e.g. correlation of say 0.3. I am not sure exactly how to do this but perhaps this might do it? $OMEGA 1 ; CL $OMEGA 1 ; V $OMEGA .09 FIXED ; COV CL=THETA(1)*EXP(ETA(1)) V=THETA(2)*EXP(ETA(3)*ETA(1)) Nick -- Nick Holford, Divn Pharmacology & Clinical Pharmacology University of Auckland, 85 Park Rd, Private Bag 92019, Auckland, New Zealand email:n.holford@auckland.ac.nz tel:+64(9)373-7599x6730 fax:373-7556 http://www.phm.auckland.ac.nz/Staff/NHolford/nholford.htm