RE: FW: OMEGA HAS A NONZERO BLOCK

From:"Steve Duffull" Subject:RE: FW: [NMusers] OMEGA HAS A NONZERO BLOCK Date:Mon, 7 Oct 2002 08:48:06 +1000 Hi all Just a quick note about my ill-conditioned VC matrix. As Ken pointed out there was high correlation between parameter 1 and 4. However knowledge of this does not solve the problem. Recall the OMEGA BLOCK(4) was $OMEGA BLOCK(4) ;BSV 0.329 0.205 0.272 0.161 -0.217 0.774 0.00818 0.0048 0.00462 0.000204 When I tried all the different permutations on this matrix it turns out the bad egg was 0.161. Using matrix banding: This works fine: $OMEGA BLOCK(4) ;BSV 0.329 0.205 0.272 0 -0.217 0.774 0 0 0.00462 0.000204 But this doesn't. $OMEGA BLOCK(4) ;BSV 0.329 0.205 0.272 0.161 -0.217 0.774 X 0.0048 0.00462 0.000204 where X is either zero or fixed at any particular (legal) value you like. This is where I'm having problems. $OMEGA BLOCK(4) ;BSV 0.329 0.205 0.272 X -0.217 0.774 0.0048 0.0048 0.00462 0.000204 I could find no values of X that allowed NONMEM to work with this matrix. Regards Steve =================================== Stephen Duffull School of Pharmacy University of Queensland Brisbane 4072 Australia Tel +61 7 3365 8808 Fax +61 7 3365 1688 http://www.uq.edu.au/pharmacy/duffull.htm University Provider Number: 00025B