[ENH] Poisson distribution
This PR adds a Poisson distribution.
The implementation is rudimentary, aiming at the minimum methods to be implemented.
The intention is to use this example to develop framework capability for discrete and mixed distributions.
FYI @julian-fong, @shreesham07, @nilesh05apr, this should be useful for all of you?
Apologies if you've already started implementing something, @nilesh05apr, I just thought it would be useful to see if any framework issues arise, since this would be the first distribution with a discrete subdomain of the reals.
I've also opened a design issue here, regarding handling of pdf vs pmf: https://github.com/sktime/skpro/issues/229
Yes, this distribution should be helpful when implmenting _predict_proba for future estimators
reviews are appreciated 😃