Posted by Marc David on April 09, 1998 at 18:11:48:
I have a set of curves that seem like they
could be described as a series of splines.
I am trying to figure out a way to take this
data and convert it into the best fitting
set of cubic splines where the total coeficients
would be less then the total data points. For
example If I had twenty data points I'd want to
describe it less then 5 cubic splines. The fewer
the better that would maximize my compression.