I'm writing a mini-library in C++ to find a 4th order Chebyshev polynomial (of the first kind) curve fit on set of x/y points varying in size (between 5-36 sets of points).
I have found a pretty good example in Python's numpy library in the numpy.polynomial.chebyshev module. See function 'chebfit' here.
I'm trying to find a paper or explanation of this method as the library is a bit brief and utilizes a lot of innate features of python that are difficult to convert to C++. This is the closest I've come to an explanation of the chebyshev.py methodology, however, there's a lot of extra stuff that's making it tricky for me to follow.
Does anybody know a better resource for following the numpy library least squares methodology, or something similar?
Thanks