PDL::Polynomial3pm
Routines for fitting with polynomials
Usage
$x = sequence(20)-10;$coeff_orig = cdouble(30,-2,3,-2); # order used in this module$y = 30-2*$x+3*$x**2-2*$x**3; # or: $y = polyval($coeff_orig->slice("-1:0"), $x->r2C);$y += ($x->grandom - 0.5) * 100;($yfit, $coeff) = fitpoly1d($x,$y,4);use PDL::Graphics::Simple;$w = pgswin();$xi = zeroes(100)->xlinvals(-10,9);$yi = polyval($coeff->r2C->slice("-1:0"), $xi->r2C);$w->plot(with=>'points',$x,$y,with=>'points',$x,$yfit, with=>'line',$xi,$yi);$yfit = fitpoly1d $data,2; # Least-squares line fit($yfit, $coeffs) = fitpoly1d $x, $y, 4; # Fit a cubic
Overview
This module contains routines for doing simple polynomial fits to data
Functions
fitpoly1d
Fit 1D polynomials to data using min chi^2 (least squares)
Usage: ($yfit, [$coeffs]) = fitpoly1d [$xdata], $data, $order, [Options...] Signature: (x(n); y(n); [o]yfit(n); [o]coeffs(order))Uses a standard matrix inversion method to do a least squares/min chi^2 polynomial fit to data. Order=2 is a linear fit (two parameters).
Returns the fitted data and optionally the coefficients (in ascending order of degree, unlike "polyval" in PDL::Math).
One can broadcast over extra dimensions to do multiple fits (except the order can not be broadcasted over - i.e. it must be one fixed scalar number like "4").
The data is normalised internally to avoid overflows (using the mean of the abs value) which are common in large polynomial series but the returned fit, coeffs are in unnormalised units.
$yfit = fitpoly1d $data,2; # Least-squares line fit($yfit, $coeffs) = fitpoly1d $x, $y, 4; # Fit a cubic$fitimage = fitpoly1d $image,3 # Fit a quadratic to each row of an image$myfit = fitpoly1d $line, 2, {Weights => $w}; # Weighted fit
Options:Weights Weights to use in fit, e.g. 1/$sigma**2 (default=1)
Releases
13 releases since Debian 2.2 (Potato); the text changed 12 times.