Hello,
I am using the function bootstrap() in S-Plus 2000 in order to get
bootstrap confidence limits for regression coefficients in a linear
model. An example of the code is following:
bootstrap(name,coef(lm(y~x1+x2+x3, data=name, na.actio=na.exclude)), B=1000)
The goal is to use the bootstrap to obtain standard error estimates and
confidence limits for the linear regression coefficients from the model
of y regressed on x1, x2 and x3.
Does it resample from the residuals of the linear model or from the
original response y?
Thank you very much,
Inma
--
Inmaculada Arostegui Madariaga
Euskal Herriko Unibertsitatea UPV/EHU
Matematika Aplikatua eta Estatistika eta Ikerkuntz Operatiboa Saila
644 Postakutxa
48080 BILBAO
Tfno: 34 946015346
Fax: 34 946013500
e-mail: inmaculada.arostegui@ehu.es
|