| Problem with SVDFIT? [message #4521] |
Tue, 13 June 1995 00:00 |
siri
Messages: 2
Registered: June 1995
|
Junior Member |
|
|
I have encountered a strange behavior of SVDFIT... I tried to do an LLS fit of a
polynomial of degree 1 (straight line) to a bunch of data points (x values
approx. 1E-9, y=c/x within [0,120]). The resulting fit seems to have the
correct slope, but the offset is completely wrong (way too small).
POLY_FIT seems to work OK, and SVDFIT works fine, if the data are scaled to
larger values. Do I have a precision problem? Can I fix it? Any ideas anybody?
And yes, I know that a straight line will not characterize my data too well, but
SVDFIT should work correctly despite that fact... :(
Thanks,
Siri
____________________________________________________________ ___________________
Siri J. Flocke Email: siri@ucar.edu
NCAR/ACD Voice: +1 (303) 497 1865
1850 Table Mesa Drive Fax: +1 (303) 497 1400
Boulder, CO. 80303
____________________________________________________________ ___________________
|
|
|
|
comp.lang.idl-pvwave archive
Members
Search
Help
Login
Home
