3 ms·
Numerical analysts are definitely familiar. Library authors may not be unless they took a course in function approximation. Personally I prefer chebyshev appro
by nimish 5y ago
Numerical analysts are definitely familiar. Library authors may not be unless they took a course in function approximation.
Personally I prefer chebyshev approximants like chebfun but pade are really good at the cost of a division.
- stephencanon 5y agoA math library implementor will generally be familiar with at least minimax and Chebyshev approximations, and generally Padé and Carathéodory-Fejér approximations as well. (Source: I implemented math libraries for a decade. We used all of these frequently.)