Re: Looking for source for normsinv function
From: Koos (ksneek_at_feweb.vu.nl)
Date: 12/22/04
- Next message: Mack: "Testing sparse binned data"
- Previous message: Phil Sherrod: "Algorithm to compute critical values of Durbin-Watson statistic?"
- In reply to: Peter J. Acklam: "Re: Looking for source for normsinv function"
- Next in thread: G Robin Edwards: "Re: Looking for source for normsinv function"
- Reply: G Robin Edwards: "Re: Looking for source for normsinv function"
- Messages sorted by: [ date ] [ thread ]
Date: Wed, 22 Dec 2004 16:59:55 +0000 (UTC)
On Sun, 19 Dec 2004 22:28:37 +0100, Peter J. Acklam wrote:
Looking at a very short code, you might try
Koos
>"Roland" <roland@nospam> wrote:
>>
>> <a
href="http://lib.stat.cmu.edu/apstat/241">http://lib.stat.cmu.edu/apstat/241>
(note: there are 2 algorithms
>> in this page, PPND7 and PPND16)
>
>"MDang" <MaiDangabc@bigpond.com.au> wrote:
>
>> Peter J. Acklam has found a more accurate algorithm :
>>
>> <a
href="http://home.online.no/~pjacklam/notes/invnorm/index.html">http://home.online.no/~pjacklam/notes/invnorm/index.html>
>
>My algorithm has a relative error in the output < 1.15e-9 which
>makes it more accurate than PPND7, but less accurate than PPND16.
>
>On the other hand, my algorithm is totally free so the algorithm
>and all implementations of it on the web page can be used without
>worrying about copyright etc.
>
>I am working on more accurate algorithms, but I don't know when
>I'll be finished.
>
>Peter
>
>--
>Peter J. Acklam - pjacklam@online.no - <a
href="http://home.online.no/~pjacklam">http://home.online.no/~pjacklam>
z = 4.91*(p.^0.14-(1-p).^0.14);
Relevant Pages
|
