PDA

View Full Version : Responses to question about H. Woltring's GCVSPL



ecross88
06-16-1994, 02:14 AM
Dear Subscribers,
A few days ago I wrote asking for some background on Herman
Woltring's GCVSPL program. I would like to thank everyone who
responded. I figured a few others might be interested in the
responses as well so here goes.
In the last mail Ellen C Ross said:
> I have been hearing alot about "Herman Woltring's classic GCVSPL fortran
> program". Can anyone give me a brief idea of what it is, how I can get
> more information, and what "GCVSPL" stands for? I also understand it is
> archived someplace. Can anyone tell me where? Is it licensed, free or
> shareware?
//////////////////////////////////////////////////////////////////////////// .
Hi Ellen,
This algorithm has been published in the past from Herman (I don't remember
the paper but I can find it if you want). Generaly itis an automatic method
for data smoothing and differentiation. It is commonly known as "Cross
Validation Splines". It is free to get it and use it. It is written in
Fortran
77 and I have the algorith on my disk. Currently I am trying to make this
file
executable for any raw data file (Because Vaughan's data are inside the
program allready). Since I program in Turbo Pascal 6 it is difficult for me
to manage it. If you know Fortran it could be easy for you.
Is it possible to e'mail me all the information that will be send to you
the next days? It is in my interest to know something more....
If you want anything else contact me again
Giannis Giakas
ggiakas@liverpool.ac.uk
////////////////////////////////////////////////////////////////////////////
Dear Ellen,
I believe I found that program just yesterday by ftp to
magnus.acs.ohio-state.edu
The program is in the pub/gaitlab directory. Sorry I don't
have any information about the program itself or what the letters
stand for. I just happened upon it by accident. Hope this helps!
Julie
JCOMMAGER@magnum.mco.edu
////////////////////////////////////////////////////////////////////////////
Dear Ellen,
Ton van den Bogert ( bogert@acs.ucalgary.ca ) implemented
this spline function in our software to analyze the gaits of horses using
the CODA-3 kinematic gait analysis system. {Adv. Eng. Software (1986)
8 (2): 104-113.}
Please contact Ton for software details. By the way: it works perfect to
calculare 1st and 2nd derivative of kinematic data !
Willem Back
Dept. of General and Large Animal Surgery,
Faculty of Veterinary Medicine,
P.O Box 80.153
3508 TD Utrecht
The Netherlands
E-mail: W.Back@hk.dgk.ruu.nl
////////////////////////////////////////////////////////////////////////////
I can tell you the following about GCVSPL:
GCVSPL is an acronym for Generalized Cross-Validatory Spline smoothing.
The program provides smoothing of kinematic data through a spline fitting
method. There are some publications by Woltring about this method,
I have been able to retrieve the following references:
An application of cross-validated splines is discussed in:
H.J. Woltring, A. de Lange, J.M.G. Kauer and R.Huiskes
"Instantaneous Helical Axis Estimation via Natural, Cross-Validated Splines"
Proc. of 5th meeting of ESB, sept. 8-10, 1986, Berlin.
In this article, reference is made to:
H.J. Woltring. Advances in Engineering Software 8(2), 104-113, 1986
(no title specified).
You can obtain the package easily from NETLIB. This is a network library
where a lot of numerical, scientific software is stored. All this software
is free and can be retrieved through FTP. I just tried to get the package
from the original FTP site of NETLIB, 'research.att.com' but my system
didn't find that. Another site that you can try, which worked for me,
is 'nac.no'.
To get GCVSPL, act as follows:
Type: 'FTP nac.no'
login as 'anonymous'
give your e-mail address as password,
go to directory 'netlib/gcv'
type: 'get gcvspl'
*** that's all ***(just type 'bye' or 'quit' to exit).
With kind regards,
Jan Cees Sabel
sabel@duttncb.tn.tudelft.nl
Delft University of Technology
The Netherlands.
//////////////////////////////////////////////////////////////////////

Thanks again for the responses,
Ellen