There are several places where the performance of aampi and stumpi can be improved via NumPy vectorized functions. Currently, we follow pseudocode from the published work which formulates the algorithm via inefficient for-loops. This isn't wrong but there are some speed enhancements that can be had by using NumPy instead.
There are several places where the performance of
aampiandstumpican be improved via NumPy vectorized functions. Currently, we follow pseudocode from the published work which formulates the algorithm via inefficient for-loops. This isn't wrong but there are some speed enhancements that can be had by using NumPy instead.