2D Polynomial Approximation
2D Polynomial approximation creates a f(x,y)=z polynomial that bests fits the
given data points. By default, all data points are included, but theoretically,
complete data is not a requirement. The result comes from a coeffecient matrix created by
calculating the given powers of x and y from indeci data for each coeffecient term,
then calculating the least square error while generating an output vector. This
variant utilizes constant Q based MFCC's for the vertical dimension.
This feature was published in August of 2010 in a McEnnis log journal entry.
