Closed wd60622 closed 8 months ago
You don't have to implement the logp or logcdf for this one, PyMC can figure it out
@wd60622 We also need to revisit the Skellam for the case rv_size_is_none
right?
@wd60622 We also need to revisit the Skellam for the case
rv_size_is_none
right?
I added it to the BNB PR. I just tagged you
Just need to solve the conflicts from merging the other PR
Thanks @wd60622 !
Have the logp tests working like the other two distributions. I tried to implement the logcdf but am hitting some difficulties with those. The values are relatively close but are off by only a decimal in a few cases. I'm checking the translation from wikipedia formula and will look for some simplifications if possible. I see the scipy implements like this. If anything is obvious, let me know