A high level script for fitting data in table files. The basic module that this script runs to do the fitting is curve_fit. You can plot the data points and the fitted curve(s) with xyplotter_auto.
xyfitter --help To get a list of available curves: % curve_fit LAs of Nov2018 this sript is still in development. I will document the details of new features in the online help document.
In some cases, particualrly in image processing tasks, the X values can be numerically large numbers (e.g. row and column values). Using the the curve_fit directly on these large numbers X values will result in severe round-off error. Hence, it is often advisable to scale the X values to a smaller range before fitting. The details of this procedure are decribe in the documention for normit.sh. This normalization is invoked by xyfitter when the local file named "X.normit" is present. The only data in the X.normit file are the values of the a0,a1 transformation parameters. The a0 term is a zeropoint, and a1 is a scale. For example, if we want the row numbers (Y values) for an image of size 1024x2048 to be convert to a range between 0 and 1, we could use values of (a0,a1) = 0.0, 2048.0.
Example or fitting polynomials.