Path: csiph.com!feeder.erje.net!2.eu.feeder.erje.net!eternal-september.org!feeder.eternal-september.org!reader02.eternal-september.org!.POSTED!not-for-mail
From: Daniele Futtorovic <da.futt.news@laposte.net>
Newsgroups: comp.lang.java.programmer
Subject: Re: How to identify outliers
Date: Tue, 28 Nov 2017 17:33:28 +0100
Organization: A noiseless patient Spider
Lines: 33
Message-ID: <ovk34t$rob$1@dont-email.me>
References: <ovjsgk$91k$1@dont-email.me> <ovjuf5$mr8$1@dont-email.me>
Mime-Version: 1.0
Content-Type: text/plain; charset=utf-8
Content-Transfer-Encoding: 8bit
Injection-Date: Tue, 28 Nov 2017 16:33:33 -0000 (UTC)
Injection-Info: reader02.eternal-september.org; posting-host="db9879cdf7deb0203cd054848c3f1d8a"; logging-data="28427"; mail-complaints-to="abuse@eternal-september.org";	posting-account="U2FsdGVkX195ktxF2m+tBdhUX+x1q58W"
User-Agent: Mozilla/5.0 (Windows NT 6.1; WOW64; rv:52.0) Gecko/20100101 Thunderbird/52.4.0
In-Reply-To: <ovjuf5$mr8$1@dont-email.me>
X-Antivirus-Status: Clean
Content-Language: en-GB
X-Antivirus: AVG (VPS 171128-0, 11/28/2017), Outbound message
Cancel-Lock: sha1:A7LiZmIlmuoJkSDzfx+Cm0jM03E=
Xref: csiph.com comp.lang.java.programmer:37702

On 2017-11-28 16:13, Eric Sosman wrote:
> 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.
> 

Spiffy. Exactly what I was looking for. I'll go over it and get back to
you if needed and you don't mind.

Thanks a ton.

-- 
DF.