| x error bars in Curvefit / MPFIT [message #44482] |
Wed, 15 June 2005 11:25 |
Sean Davis
Messages: 19
Registered: August 1999
|
Junior Member |
|
|
Hello All,
First, I apologize if this topic has been brought up before (I searched
google groups and couldn't find a post on it).
I have some lab calibration data (x,y pairs, say), and I am trying to fit y
as a function of x using the Markwardt fitting library (I suppose the same
question could apply if I were using CURVEFIT, too).
In my fit, x represents a "raw" value read from an instrument (it's actually
calculated from several other "raw" values), and y represents the "true"
value from the calibration using a different instrument. The goal is to
come up with a fit that can be applied to my raw data that best fits the
"truth".
For fitting using MPFITFUN or similar programs, one typically provides error
bars in the y value. But in my case, I have known errors in both x and y.
I know that in the simple case of fitting x and y to a line, errors in x can
be expressed as errors in y, but is there an easier way to do this. The
conversion of errors in x to errors in y does not seem very straightforward
for complicated fits.
Thanks
|
|
|
|
comp.lang.idl-pvwave archive
Members
Search
Help
Login
Home
