ortho_fit.pro

NRAO GBTIDL
Produced by IDL 6.1
Developer Documentation

prev file | next file view single page | view frames | update routines

SUMMARY: fields | routine DETAILS: routine

./toolbox
ortho_fit.pro

Last modification date:: Thu Jun 2 23:56:08 2005

ortho_fit

function ortho_fit(xx, yy, nfit, cfit, rms)

Function uses general orthogonal polynomial to do least squares fitting.

This code came from Tom Bania's GBT_IDL work. Local modifications include:

Documentation modified for use by idldoc

Array syntax changed to [] and compile_opt idl2 used.

Indententation used to improve readability.

Unnecessary code removed.

Some argument checks added.

Returns

polyfit array: contains information for fitting to arbitrary points using the recursion relations

Examples

fit a 3rd order polynomial to the data in !g.s[0]

    yy = *(!g.s[0].data_ptr)
    xx = dindgen(n_elements(yy))
    f = ortho_fit(xx, yy, 3, cfit, rms)

Version

$Id: ortho_fit.pro,v 1.2 2004/12/08 03:50:36 bgarwood Exp $

Parameters

xx
in, required
The x-values to use in the fit.

yy
in, required
The data to be fit at xx.

nfit
in, required
integer The order of the polynomial to fit.

cfit
out, required
variable On return, cfit has coefficients of the polynomial
fit = sum(m) a_m x^m
because of round-off using this form gives unreliable results above about order 15 even for double precision. See ortho_poly for a better discussion on the contents of cfit.

rms
out, required
variable The rms error for each polynomial up to order nfit.

Produced by IDLdoc 1.6 on Thu Jun 2 23:56:22 2005

Parameters
xx in, required	The x-values to use in the fit.
yy in, required	The data to be fit at xx.
nfit in, required integer	The order of the polynomial to fit.
cfit out, required variable	On return, cfit has coefficients of the polynomial fit = sum(m) a_m x^m because of round-off using this form gives unreliable results above about order 15 even for double precision. See ortho_poly for a better discussion on the contents of cfit.
rms out, required variable	The rms error for each polynomial up to order nfit.