[R] MLE for probit regression. How to avoid p=1 or p=0

[Rui Barradas]

Hello,

You write a function of two arguments, 'par' and 'data' and do not use 
them in the body of the function. Furthermore, what are b0, b1x and y?

Also, take a look at ?.Machine. In particular, couldn't you use

precision0 <- .Machine$double.eps
precision1 <- 1 - .Machine$double.eps

instead of 0.000001 and 0.999999?

Hope this helps,

Rui Barradas

Em 27-05-2013 16:21, knouri escreveu:
>
>
> Dear all:
>
> I am writing the following small function for a probit likelihood.
> As indicated, in order to avoid p=1 or p=0, I defined some precisions.
> I feel however, that there might be a better way to do this.
> Any help is greatly appreciated.
>
> ##########################################################
>
> ##set limits to avoid px=0 or px=1
> precision1   <- 0.999999
> precision0   <- 0.000001
>
> logpost <- function(par, data){
> px    <- pnorm(b0 + b1x)
> # to avoid px=1 or px=0
> px[px >  precision1] <- precision1
> px[px <  precision0] <- precision0
> loga  <- sum( y*log(px)+(1-y)*log(1-px) )
> loga
> }
>
> #########################################################
>
>
>
> Best,
>
>
> ________________________________
> Keramat Nourijelyani, PhD
> Associate Professorof Biostatistics
> Tehran University of Medical Sciences
> http://tums.ac.ir/faculties/nourij
>
> 	[[alternative HTML version deleted]]
>
>
>
> ______________________________________________
> R-help at r-project.org 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.
>