pspp-dev
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

[patch #6635] Wilcoxon signed rank test


From: John Darrington
Subject: [patch #6635] Wilcoxon signed rank test
Date: Thu, 18 Sep 2008 11:43:59 +0000
User-agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.9.0.1) Gecko/2008071618 Iceweasel/3.0.1 (Debian-3.0.1-1)

URL:
  <http://savannah.gnu.org/patch/?6635>

                 Summary: Wilcoxon signed rank test
                 Project: PSPP
            Submitted by: jmd
            Submitted on: Thu 18 Sep 2008 07:43:56 PM WST
                Category: None
              Item Group: None
                  Status: Ready For Test/Review
             Assigned to: None
        Originator Email: 
             Open/Closed: Open
         Discussion Lock: Any

    _______________________________________________________

Details:

This patch adds the /WILCOXON subcommand to NPAR TESTS. 

Some things to note:

1. The other software produces a value titled "point probablility".  I've no
idea what this means or how to generate it, so I've left it out.  If somebody
knows what it means and how to calculate it, then I'll include it.

2. I've to calculate the wilcoxon significance, I've used a routine from Dr.
Rob van Son whom I contacted, and he's kindly agreed to put this code under
GPLv2+  The results from this are slightly different to those from the Chicago
sofware --- about 1/2 digits in the second decimal place.  Looking at their
documentation, I infer that they're using some very anachronistic routine
(they talk about running out of memory when calculating it), so frankly I'm
more inclined to trust this routine than the other one.

3. Like the other software, the /METHOD=EXACT option takes a rediculously
long time, if the number of cases is more than about 20.  Therefore the TIMER
option sets a maximum time to try.  Although I thought it would be easy
pre-empt a routine
on a timer in posix, it wasn't so simple after all.  I ended up jumping out
of a signal handler with a longjmp --- somewhat messy, but after some thought,
I can't see any reason why it's not portable or not safe (on a single threaded
posix system).





    _______________________________________________________

File Attachments:


-------------------------------------------------------
Date: Thu 18 Sep 2008 07:43:56 PM WST  Name: wilcoxon.patch  Size: 39kB   By:
jmd

<http://savannah.gnu.org/patch/download.php?file_id=16518>

    _______________________________________________________

Reply to this item at:

  <http://savannah.gnu.org/patch/?6635>

_______________________________________________
  Message sent via/by Savannah
  http://savannah.gnu.org/





reply via email to

[Prev in Thread] Current Thread [Next in Thread]