[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: fitting functions which contain 'i'
From: |
Przemek Klosowski |
Subject: |
Re: fitting functions which contain 'i' |
Date: |
Thu, 5 Feb 2009 16:12:09 -0500 (EST) |
Does Octave have any ability to fit a function which contains 'i' (the
square root of -1 in case 'i' is as unclear on your screen as it is on
mine). The least squares fitting tool does not seem to work, or at least
I could not get to work, and these are two quite different things, of
course.=20
I know the feeling :) Which least squares fitting tool did you use? More
importantly, how are you using the complex numbers---is your fitted function
complex-valued? That can be seen as two-dimensional fitting, so you can
convert it to a single-dimensional figure of merit by taking the sum of squares
of deviations (possibly divided by errors), i.e. the standard chi-square
metric. If this is not appropriate for your data, you need to come up with
another single-dimensional figure of merit, because it is impossible to optimize
a 2-dimensional quantity---there is no strict ordering in a complex plane.
Please describe more about your data and fitting requirements.