[R] Non-Central t

Fri May 18 17:24:00 CEST 2001

"Morris, Jeffrey [OCDUS]" <JMORRIS6 at OCDUS.JNJ.COM> writes:

> In the help file for the non-central t, the following appears:
> 
> ncp: non-centrality parameter delta; currently `ncp <= 37.62'.
> 
> I assume that this means the ncp cannot exceed 37.62.  Is this
> still the case and is there any plans to increase this restriction?
> 
> Thanks!

Judging from the comments in the code, the problem is still there.

Eventually, we want to get rid of these restrictions, and I recently
did so for the noncentral chisquare density, but it requires nontrivial
changes to the algorithms and can be tricky to get right (as witnessed
by the subsequent bugfix to dnchisq.c...). A typical problem is that
the noncentral distributions involve infinite sums where the terms
satisfy a multiplicative recurrence relation, but where extreme
arguments cause the recurrence to start from zero due to floating
point underflow.

> 
> ------_=_NextPart_001_01C0DFA2.5F9BB5B0
> Content-Type: text/html
> 
> <!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 3.2//EN">
> <HTML>
> <HEAD>

 *sigh*

-- 
   O__  ---- Peter Dalgaard             Blegdamsvej 3  
  c/ /'_ --- Dept. of Biostatistics     2200 Cph. N   
 (*) \(*) -- University of Copenhagen   Denmark      Ph: (+45) 35327918
~~~~~~~~~~ - (p.dalgaard at biostat.ku.dk)             FAX: (+45) 35327907
-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-
r-help mailing list -- Read http://www.ci.tuwien.ac.at/~hornik/R/R-FAQ.html
Send "info", "help", or "[un]subscribe"
(in the "body", not the subject !)  To: r-help-request at stat.math.ethz.ch
_._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._