Re: OMEGA HAS A NONZERO BLOCK

From:"Lewis B. Sheiner" Subject:Re: [NMusers] OMEGA HAS A NONZERO BLOCK Date:Mon, 07 Oct 2002 09:19:23 -0700 All, My only quibble with Ken's reply, which I tried to make clear in my original note, is that to 'simplify the model' (from a mechanistically reasonable more complex one to an 'adequate' less complex one) is FORMALLY IDENTICAL to fixing certain parameters (of the more complex model) to sharp values (usually on their boundary), and hence, to adding additional information. You just can't escape (as Ken says, " you gotta live with what you've got"): To make progress, you've often got to assert more than your data can support all by themselves. If your only goal is describing the data at hand, or testing hypotheses to which the data at hand speak, then by all means, use the simplest model supported by the data which affords reasonable power. But if your goal is a model useful for extrapolation (prediction in new circumstances), control, or design, then such a strict empiricist policy can lead to disaster. For predictive models, the inadequacy of the data at hand means you must resort to science, and in that case, the domain experts (or regulators) get to have the last word on what constitutes valid 'extra information'. I repeat my outrageous claim: You never know nothing. LBS. _/ _/ _/_/ _/_/_/ _/_/_/ Lewis B Sheiner, MD (lewis@c255.ucsf.edu) _/ _/ _/ _/ _/ Professor: Lab. Med., Biophmct. Sci. _/ _/ _/ _/_/_/ _/_/ Mail: Box 0626, UCSF, SF,CA,94143 _/ _/ _/ _/ _/ Courier: Rm C255, 521 Parnassus,SF,CA,94122 _/_/ _/_/ _/_/_/ _/ 415-476-1965 (v), 415-476-2796 (fax)