Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]

Groups > comp.lang.java.programmer > #37700

Re: How to identify outliers

Show all headers | View raw

On 11/28/2017 9:40 AM, Daniele Futtorovic wrote:
> Hi folks,
> 
> I got a statistical problem, and due to a youth math trauma, I'd like a
> bit of help to solve it.
> 
> I have a data series. Integer values, but that should be beside the
> point. I want to identify and eliminate "outliers" from this series.
> "Outliers" being values that are "too big" or "too small", in short,
> that are not congruent with the rest of the series. I am well aware that
> saying "outlier" begs the question, hence the quotes, but I can
> fine-tune the degree of outlying for my purposes later on. For the time
> being, I wonder if someone could outline some basic approaches I could
> take in the matter.


http://newton.newhaven.edu/sross/piercescriterion.pdf

     If the table at the end of this (short) paper isn't suited to
your needs, I can also offer source code (in C) to adapted from
https://groups.google.com/d/topic/sci.math/Kypai3jqss8/discussion
to calculate the underlying distribution.  Drop me a line if you'd
like it.

-- 
esosman@comcast-dot-net.invalid
Eleven hundred forty-nine days to go.

Back to comp.lang.java.programmer | Previous | Next — Previous in thread | Next in thread | Find similar | Unroll thread

Thread

How to identify outliers Daniele Futtorovic <da.futt.news@laposte.net> - 2017-11-28 15:40 +0100
  Re: How to identify outliers Eric Sosman <esosman@comcast-dot-net.invalid> - 2017-11-28 10:13 -0500
    Re: How to identify outliers Daniele Futtorovic <da.futt.news@laposte.net> - 2017-11-28 17:33 +0100
  Re: How to identify outliers Eric Douglas <e.d.programmer@gmail.com> - 2017-11-28 07:28 -0800
    Re: How to identify outliers Daniele Futtorovic <da.futt.news@laposte.net> - 2017-11-28 17:35 +0100
  Re: How to identify outliers Robert Klemme <shortcutter@googlemail.com> - 2017-11-30 18:38 +0100

csiph-web