Jump to content

Talk:Freedman–Diaconis rule

Page contents not supported in other languages.
From Wikipedia, the free encyclopedia

isn't it binSize = 2 * IQR / N^(1/3) ?? —Preceding unsigned comment added by LN29 (talkcontribs)

That's what the article currently says. Michael Hardy (talk) 20:13, 8 November 2009 (UTC)[reply]
It's also what it said when LN29 posted this comment. Michael Hardy (talk) 20:14, 8 November 2009 (UTC)[reply]

wrong formula

[edit]

The formula as stated doesn't give anything remotely correct. Looking around on the web a formula that looks right and actually seems to work in my tests is:

(N^1/3 * range)/(2 * IQR)

Where range is the range of the data. —Preceding unsigned comment added by 64.157.93.210 (talkcontribs)


The formula listed is the formula found in the paper. Could you maybe be thinking of the number of bins rather than the bin size? (Lucas(CA) (talk) 19:07, 5 October 2016 (UTC))[reply]