Feature expectations approach #3
Description
To clarify, your method assumes the input moments correspond to the traditional first k moments ie. x, x^2... x^k of a distribution, correct? What if I would like to find the maximum entropy distribution for a custom set of feature expectations?