I wonder whether Padé approximants are well known by this kind of researcher. E.g. http://www-labs.iro.umontreal.ca/~mignotte/IFT2425/Documents...
Personally I prefer chebyshev approximants like chebfun but pade are really good at the cost of a division.