>>>>> "GB" == Göran Broström <[EMAIL PROTECTED]>
>>>>> on Thu, 26 Dec 2002 01:17:21 +0100 (CET) writes:
GB> On 25 Dec 2002, Peter Dalgaard BSA wrote:
>> [EMAIL PROTECTED] writes:
>>
>> > The problem is caused by exp(x) returning Inf when x is
>> large.
>> >
>> > One way of fixing the problem [there may be better
>> ways] would be to include the > line
>> >
>> > out[x > 709] <- 1
>> >
>> > in inv.logit()
>>
>> Yes, or replace the entire code with something like
>>
>> inv.logit2 <- function(x) ifelse(x > 0, 1 - 1/(1+exp(x)),
>> 1/(1+exp(-x)))
GB> Or use 'plogis(x)', eventually with 'lower.tail = ...'
GB> ...
yes, really!
inv.logit() should never have made its way into the boot
package at all I think.
I mean neither in tto the S-plus "library". plogis() has bveen
part of S since ages AFAIK.
Martin Maechler <[EMAIL PROTECTED]> http://stat.ethz.ch/~maechler/
Seminar fuer Statistik, ETH-Zentrum LEO C16 Leonhardstr. 27
ETH (Federal Inst. Technology) 8092 Zurich SWITZERLAND
phone: x-41-1-632-3408 fax: ...-1228 <><
______________________________________________
[EMAIL PROTECTED] mailing list
http://www.stat.math.ethz.ch/mailman/listinfo/r-devel
