[R] Optim: Function definition

From: Serguei Kaniovski <kaniovsk_at_wifo.ac.at>
Date: Wed 04 Oct 2006 - 13:09:04 GMT


Hi all,

I apply "optim" to the function "obj", which minimizes the goodness of fit statistic and obtains Pearson minimum chi-squared estimate for x[1], x[2] and x[3]. The vector "fr" contains the four observed frequencies.

Since "fr[i]" appears in the denominator, I would like to substitute "0" in the sum if fr[i]=0.

I tried an "ifelse" condition which works, but in some cases the solution seems rather odd. Is there anything wrong with the way second objective function is coded?

obj<-function(x){
(fr[1]-x[1]*x[2]-x[3]*sqrt(x[1]*(1-x[1])*x[2]*(1-x[2])))^2/fr[1]+
(fr[2]-(1-x[1])*x[2]+x[3]*sqrt(x[1]*(1-x[1])*x[2]*(1-x[2])))^2/fr[2]+
(fr[3]-x[1]*(1-x[2])+x[3]*sqrt(x[1]*(1-x[1])*x[2]*(1-x[2])))^2/fr[3]+
(fr[4]-(1-x[1])*(1-x[2])-x[3]*sqrt(x[1]*(1-x[1])*x[2]*(1-x[2])))^2/fr[4]
}

obj<-function(x){

ifelse(fr[1]==0,0,(fr[1]-x[1]*x[2]-x[3]*sqrt(x[1]*(1-x[1])*x[2]*(1-x[2])))^2/fr[1])+
ifelse(fr[2]==0,0,(fr[2]-(1-x[1])*x[2]+x[3]*sqrt(x[1]*(1-x[1])*x[2]*(1-x[2])))^2/fr[2])+
ifelse(fr[3]==0,0,(fr[3]-x[1]*(1-x[2])+x[3]*sqrt(x[1]*(1-x[1])*x[2]*(1-x[2])))^2/fr[3])+
ifelse(fr[4]==0,0,(fr[4]-(1-x[1])*(1-x[2])-x[3]*sqrt(x[1]*(1-x[1])*x[2]*(1-x[2])))^2/fr[4])
}

sval=rep(0.1,3)
fr<-c(0,0.1,0.2,0.3)

optim(sval,obj, method="BFGS")$par

Thank you,
Serguei

-- 
___________________________________________________________________

Austrian Institute of Economic Research (WIFO)

Name: Serguei Kaniovski			P.O.Box 91
Tel.: +43-1-7982601-231			Arsenal Objekt 20
Fax:  +43-1-7989386			1103 Vienna, Austria
Mail: Serguei.Kaniovski@wifo.ac.at

http://www.wifo.ac.at/Serguei.Kaniovski

______________________________________________
R-help@stat.math.ethz.ch mailing list
https://stat.ethz.ch/mailman/listinfo/r-help
PLEASE do read the posting guide http://www.R-project.org/posting-guide.html
and provide commented, minimal, self-contained, reproducible code.
Received on Wed Oct 04 23:11:38 2006

Archive maintained by Robert King, hosted by the discipline of statistics at the University of Newcastle, Australia.
Archive generated by hypermail 2.1.8, at Sat 14 Oct 2006 - 22:30:10 GMT.

Mailing list information is available at https://stat.ethz.ch/mailman/listinfo/r-help. Please read the posting guide before posting to the list.