|
|
well I was meaning that I have some sort
of functions like sin or cos, more generally a function f(x,y,z,...)
I have some points in a space and I would try to approximate them
using a finite number of operators.
I have seen some documents discussing this subject.
They use exactly operator(a,b): a*sin(x*b) = y
where x, y are the coordinates of the point and
a, b the parameters of the operator.
Actually they use multiple operators like that to approximate.
What it really misses in those documents was how to calculate
the parameters.... quite important indeed.
the other point was what happens when I fix the number of the operators
and...
of course, what happens when I fix the maximum error!
Anyway those documents (not really papers) calls it approximation by
operators...
but I was unable to find more on this subject...
any help is appreciated
GG
|
|