s-news
[Top] [All Lists]

fitting byvariate model

To: "s-plus" <s-news@wubios.wustl.edu>
Subject: fitting byvariate model
From: "Xao Ping" <xping@lycos.com>
Date: Fri, 15 Dec 2000 05:25:24 +0800
Organization: Lycos Communications (http://comm.lycos.com:80)
Reply-to: xping@lycos.com
Dear All:
I am trying to obtain analytical expression for a function of two variables
q(p,N). A have a data matrix Q[i,j]=Q[p[i],N[j]] produced my Monte Carlo 
simulation;

I found that satisfactory representation of this function may be as folows:

q[p,N]=a(p)+b(p)*N+c(p)*N^2

where
a(p)=(Aa+Ba*p)*exp(Ca*p)+(Da+Ea*p)*exp(-Fa*p)
b(p)=(Ab+Bb*p)*exp(Cb*p)+(Db+Eb*p)*exp(-Fb*p)
c(p)=(Ac+Bc*p)*exp(Cc*p)+(Dc+Ec*p)*exp(-Fc*p)

Question:

Is that possible to use some of standard S+ functions to perform this kind 
of fitting and obtain estimates for variances of 18 parameters. I would like to 
avoid sequantial determination of parameters (that's what I've already done).
I want to fit the whole thing in a single process
 

Aa,Ba,Ca,Da,Ea,Fa
Ab,Bb,Cb,Db,Eb,Fb
Ac,Bc,Cc,Dc,Ec,Fc

Thank you
Xao Ping

R&R Pharmakinetics
Taiwan


 




Get FREE Email/Voicemail with 15MB at Lycos Communications at 
http://comm.lycos.com

<Prev in Thread] Current Thread [Next in Thread>
  • fitting byvariate model, Xao Ping <=