Alpha and Beta can be out of bounds for beta distribution causing Nans to appear
lajohn4747 opened this issue · comments
John La commented
Environment Details
Please indicate the following details about the environment in which you found the bug:
- Copulas version: 0.9.2.dev1
- Python version: 3.11
- Operating System: Mac
Error Description
ppf function here returns a Nan if alpha and beta values are not within bounds (i.e. value > 0):
Copulas/copulas/univariate/base.py
Line 582 in 64a91e8
This is causing issues where Nans are appearing in synthetic data where the original data does not contain nans.
Values are being passed in from here:
https://github.com/sdv-dev/SDV/blob/71d61d6e8b01808c9b5a70b4573fdc7e1749b758/sdv/single_table/copulas.py#L345