Re: posthoc step

From: "Nick Holford" n.holford@auckland.ac.nz Subject: Re: [NMusers] posthoc step Date: Tue, December 14, 2004 3:56 pm Hi, Would it be fair to summarize the performance of the various methods (with fixed residual SD) as follows? 1. NONMEM using either the DATA or the PRIOR methods and WinBugs (Steve, is this correct?) get past the local minimum 2. NONMEM MAXEVAL=0 and SAS get stuck in the local minimum Tom, thanks for discovering the local minimum and describing it. For the information of those who (like me) can 'see' things better in graphical form this Excel worksheet http://www.health.auckland.ac.nz/pharmacology/staff/nholford/pkpd/Bayesian/BLS_solver.xls illustrates how the OBJ value changes with differential values of K ('Ktry'). The observation derived component (OLS) and the Bayesian derived component (BLS) of the OBJ are graphed to show how they combine to produce the local and presumed global minimum. If you have the Excel Solver installed you can test its performance e.g. with initial estimates of Khat of 10 and 1. Nick -- Nick Holford, Dept Pharmacology & Clinical Pharmacology University of Auckland, 85 Park Rd, Private Bag 92019, Auckland, New Zealand email:n.holford@auckland.ac.nz tel:+64(9)373-7599x86730 fax:373-7556 http://www.health.auckland.ac.nz/pharmacology/staff/nholford/