[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
PSPP-BUG: [bug #43768] Histogram in FREQUENCIES incorrect
From: |
John Darrington |
Subject: |
PSPP-BUG: [bug #43768] Histogram in FREQUENCIES incorrect |
Date: |
Sat, 06 Dec 2014 18:54:04 +0000 |
User-agent: |
Mozilla/5.0 (X11; Linux i686; rv:31.0) Gecko/20100101 Firefox/31.0 Iceweasel/31.2.0 |
Follow-up Comment #1, bug #43768 (project pspp):
I don't think it is right to say that either histogram is "incorrect". For
any dataset, there are an infinite number of possible histograms. The
difference is that they have different number of bins.
In particular, the EXAMINE command uses Sturges Rule[1] to determine the
bin_width, whereas FREQENCIES uses the Freedman-Diaconis rule[2]. The former
is generally regarded as being better for a small number of data, but not so
good for large datasets.
[1] http://www.syque.com/improvement/Sturgess%20Rule.htm
[2] https://en.wikipedia.org/wiki/Freedman%E2%80%93Diaconis_rule
_______________________________________________________
Reply to this item at:
<http://savannah.gnu.org/bugs/?43768>
_______________________________________________
Message sent via/by Savannah
http://savannah.gnu.org/