[Biodevelopers] Re: Biodevelopers digest, Vol 1 #30 - 1 msg

[Allen Henry]

I'm only a student in the BioInfo biz, but am very
good at math and come from an image processing
background where we called this the Root Mean Square
Error (RMSE). I recognized this right away because one
of my first 'bugs' that I fixed was in this very same
function. I found something on Google's groups with a
similar question as well as the equation:

http://thesaurus.maths.org/dictionary/map/word/3701

I hope this of help

:-)

=======================================================

Hi all,

Given a set of of tiepoints for images and/or maps,
how should one go about calculating the RMSE?  This is
with respect to surface fitting models for image
registration
and rectification.  Such methods include polynomials
and piecewise linear functions.  No prior assumptions
'bout anything.  This, of course, rules out kriging
for rubber-sheeting... well, not exactly, but that's a
question for another day.

So, given that this is a matter of developing a
deterministic model, would YOU account for degrees of
freedom?  Two alternatives follow below for
polynomials.

(1)
RMS for X    = sqrt ( sum( (residual X)**2 ) / (N - K)
)
RMS for Y    = sqrt ( sum( (residual Y)**2 ) / (N - K)
)
RMS Distance = sqrt ( (RMS for X)**2 + (RMS for Y)**2
)

(2)
RMS for X    = sqrt ( sum( (residual X)**2 ) / N )
RMS for Y    = sqrt ( sum( (residual Y)**2 ) / N) )
RMS Distance = sqrt ( (RMS for X)**2 + (RMS for Y)**2
)

where,

N is the number of GCPs and K is the number of terms
for the specified order.

Regards.

-David
fogel at geog.ucsb.edu


Hope this Helps
--- biodevelopers-request at bioinformatics.org wrote:
> Send Biodevelopers mailing list submissions to
> 	biodevelopers at bioinformatics.org
> 
> To subscribe or unsubscribe via the World Wide Web,
> visit
> 
>
http://bioinformatics.org/mailman/listinfo/biodevelopers
> or, via email, send a message with subject or body
> 'help' to
> 	biodevelopers-request at bioinformatics.org
> 
> You can reach the person managing the list at
> 	biodevelopers-admin at bioinformatics.org
> 
> When replying, please edit your Subject line so it
> is more specific
> than "Re: Contents of Biodevelopers digest..."
> 
> 
> Today's Topics:
> 
>    1. Re: [BiO BB] delta RMS (Joseph Landman)
> 
> --__--__--
> 
> Message: 1
> From: Joseph Landman
> <landman at scalableinformatics.com>
> To: bio_bulletin_board at bioinformatics.org
> Cc: biodevelopers <biodevelopers at bioinformatics.org>
> Date: 18 Aug 2002 22:08:34 -0400
> Subject: [Biodevelopers] Re: [BiO BB] delta RMS
> Reply-To: biodevelopers at bioinformatics.org
> 
> Hi Pete:
> 
>   RMS deviation (also known as standard deviation)
> is generally well
> defined.  Could what you have be 
> 
> 	delta (RMS deviation)[i,j] = SD(i) - SD(j)
> 
> basically using the delta as a difference operator
> between two different
> SD's?  This might be one of several possible
> "signatures" that an
> analysis would use to compare measurement
> distributions, or set
> thresholds for sub-sampling to help delineate
> clusters.
> 
> 
> On Sun, 2002-08-18 at 21:54, Peter oledzki wrote:
> > Hello,
> > 
> > I'm working on a project at the moment and I've
> come
> > across something called delta root mean squared
> > deviation....does anybody know what this is?
> > 
> > Could they possily try and explain it to me.....?
> > 
> > Any help would be much appreciated.
> > 
> > Cheers 
> > 
> > Pete Oledzki
> 
> -- 
> Joseph Landman, Ph.D
> Scalable Informatics LLC
> email: landman at scalableinformatics.com
>   web: http://scalableinformatics.com
> phone: +1 734 612 4615
> 
> 
> 
> --__--__--
> 
> _______________________________________________
> Biodevelopers mailing list
> Biodevelopers at bioinformatics.org
>
http://bioinformatics.org/mailman/listinfo/biodevelopers
> 
> 
> End of Biodevelopers Digest


=====
===================
Allen Henry
(408) 799-9404 cell
10am - 10pm

__________________________________________________
Do You Yahoo!?
HotJobs - Search Thousands of New Jobs
http://www.hotjobs.com