commons-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Kim van der Linde <kimvdli...@yahoo.com>
Subject Re: [math] Fwd: Re: Median in stats
Date Thu, 12 Aug 2004 21:25:37 GMT
I just downloaded it, so if anyone wants a copy, let
me know....

Theory and Methods

The Remedian: A Robust Averaging Method for Large Data
Sets

Peter J. Rousseeuw; Gilbert W. Bassett, Jr.

Journal of the American Statistical Association, Vol.
85, No. 409. (Mar., 1990), pp. 97-104.

Stable URL:
http://links.jstor.org/sici?sici=0162-1459%28199003%2985%3A409%3C97%3ATRARAM%3E2.0.CO%3B2-R

Abstract

It is often assumed that to compute a robust estimator
on n data values one needs at least n storage elements
(contrary to the sample average, that may be
calculated with an updating mechanism). This is one of
the main reasons why robust estimators are seldom used
for large data sets and why they are not included in
most statistical packages. We introduce a new
estimator that takes up little storage space,
investigate its statistical properties, and provide an
example on real-time curve äveraging" in a medical
context. The remedian with base b proceeds by
computing medians of groups of b observations, and
then medians of these medians, until only a single
estimate remains. This method merely needs k arrays of
size b (where n = bk), so the total storage is O(logn)
for fixed b or, alternatively, O(n 1/k ) for fixed k.
Its storage economy makes it useful for robust
estimation in large data bases, for real-time
engineering applications in which the data themselves
are not stored, and for resistant äveraging" of curves
or images. The method is equivariant for monotone
transformations. Optimal choices of b with respect to
storage and finite-sample breakdown are derived. The
remedian is shown to be a consistent estimator of the
population median, and it converges at a nonstandard
rate to a median-stable distribution.
---------

Cheers,

Kim

--- "Mark R. Diggory" <mdiggory@latte.harvard.edu>
wrote:

> This is about as close as I can get to a citation in
> JSTOR without 
> authentication:
> 
>
http://links.jstor.org/sici?sici=0162-1459%28199003%2985%3A409%3C97%3ATRARAM%3E2.0.C0%3B2-R
> 
> Here's an abstract of it I found via google.
> 
> http://www.agoras.ua.ac.be/abstract/Remrob90.htm
>
http://www.inomics.com/cgi/repec?handle=RePEc:boc:bocode:R141101
> 
> 
> Phil Steitz wrote:
> 
> >Posting a reference to the paper would be better
> ;-)
> > 
> >-Phil
> >
> >	-----Original Message----- 
> >	From: Mark R. Diggory
> [mailto:mdiggory@latte.harvard.edu] 
> >	Sent: Thu 8/12/2004 1:52 PM 
> >	To: Jakarta Commons Developers List 
> >	Cc: David Aubespin 
> >	Subject: Re: [math] Fwd: Re: Median in stats
> >	
> >	
> >
> >	Sorry about that, JSTOR just gets more and more
> anal about even
> >	searching thier contents.
> >	
> >	I'll send it to anyone privately who wishes to see
> it. -Mark
> >	
> >	Phil Steitz wrote:
> >	
> >	>I agree that this looks interesting.  The link
> below is to an authenticated site.  Can you (David)
> provide a full reference to the paper?
> >	>
> >	>Phil
> >	>
> >	>       -----Original Message-----
> >	>       From: Mark R. Diggory
> [mailto:mdiggory@latte.harvard.edu]
> >	>       Sent: Thu 8/12/2004 1:31 PM
> >	>       To: David Aubespin; Jakarta Commons
> Developers List
> >	>       Cc:
> >	>       Subject: [math] Fwd: Re: Median in stats
> >	>      
> >	>      
> >	>
> >	>       I think we would be very interested in
> such an approach. I'll forward
> >	>       this onto the list so others may know
> about it.
> >	>      
> >	>       Here's one of the authors:
> >	>      
> http://tigger.uic.edu/~gib/vit.htm#Statistics
> >	>      
> >	>       Their paper:
> >	>      
>
http://www.jstor.org.ezp2.harvard.edu/view/01621459/di985983/98p0207b/0?config=jstor&frame=noframe&userID=80673cc8@harvard.edu/018dd5533b005013e31f0&dpi=3
> >	>      
> >	>       David Aubespin wrote:
> >	>      
> >	>       > Hi,
> >	>       >
> >	>       > I have an implementation of a median
> computation based on the Remedian
> >	>       > algorithm (a paper had be published in
> 1990 about it).  Its main
> >	>       > advantage is that it takes up a fixed
> amount of memory regardless of
> >	>       > the data sample size (very useful when
> dealing with large datasets
> >	>       > like in astronomy).
> >	>       >
> >	>       > I was wondering if I could contribute it
> to the Math project.
> >	>       >
> >	>       > Let me know,
> >	>      
> >	>      
> >	>       thanks,
> >	>       Mark
> >	>      
> >	>       --
> >	>       Mark R. Diggory
> >	>       Software Developer
> >	>       Harvard MIT Data Center
> >	>       http://www.hmdc.harvard.edu
> >	>      
> >	>      
> >	>      
>
---------------------------------------------------------------------
> >	>       To unsubscribe, e-mail:
> commons-dev-unsubscribe@jakarta.apache.org
> >	>       For additional commands, e-mail:
> commons-dev-help@jakarta.apache.org
> >	>      
> >	>      
> >	>
> >	> 
> >	>
> >	
> >	
> >	--
> >	Mark R. Diggory
> >	Software Developer
> >	Harvard MIT Data Center
> >	http://www.hmdc.harvard.edu
> >	
> >	
> >
>
---------------------------------------------------------------------
> >	To unsubscribe, e-mail:
> commons-dev-unsubscribe@jakarta.apache.org
> >	For additional commands, e-mail:
> commons-dev-help@jakarta.apache.org
> >	
> >	
> >
> >  
> >
> 
> 
> -- 
> Mark R. Diggory
> Software Developer
> Harvard MIT Data Center
> http://www.hmdc.harvard.edu
> 
> 
>
---------------------------------------------------------------------
> To unsubscribe, e-mail:
> commons-dev-unsubscribe@jakarta.apache.org
> For additional commands, e-mail:
> commons-dev-help@jakarta.apache.org
> 
> 



		
__________________________________
Do you Yahoo!?
Yahoo! Mail is new and improved - Check it out!
http://promotions.yahoo.com/new_mail

---------------------------------------------------------------------
To unsubscribe, e-mail: commons-dev-unsubscribe@jakarta.apache.org
For additional commands, e-mail: commons-dev-help@jakarta.apache.org


Mime
View raw message