Add prediction type to return the mean, variance, and mode

[neverfox]

The mode is only included if the underlying distribution implements it (e.g. zero-inflated does not). Additionally, this change avoids unnecessary sampling if the prediction type doesn't require it. Raises a ValueError for invalid prediction type for Expectile distribution.

[StatMixedML]

@neverfox Thanks for the PR.

Before merging, can you please

• include try also for the mean and variance in case one or all of them are not available
• rename properties and pred_props to moments and pred_moments, also in the docstrings, since this is more understandable
• add an example notebook, e.g., for the Gaussian and include the properties in the prediction section

[StatMixedML]

@neverfox Thanks for the PR.

Before merging, can you please

• include try also for the mean and variance in case one or all of them are not available
• rename properties and pred_props to moments and pred_moments, also in the docstrings, since this is more understandable
• add an example notebook, e.g., for the Gaussian and include the properties in the prediction section

[neverfox]

No problem. I think it cleaner at this point to move the logic to a get_moments method, since it's starting to take over the body of predict_dist. Do you agree? Also, would you prefer it to just return None if there's nothing to do (as you did with draw_samples) or continue with the ValueError approach?

[StatMixedML]

• Yes, having a get_moments function that has the fitted distribution as an argument within the DistributionClass makes sense
• If one of the moments is not implemented, returning None is probably better than the ValueError way

Thanks for the suggestions.

[StatMixedML]

I was wondering that if the variance or mode are not available in the distribution class, we can also infer them from the predicted samples. Can you add this maybe.

[neverfox]

Sure. It might make sense to make that behavior explicit so the user knows what they're getting? Jul 1, 2023 8:22:19 AM Alexander März ***@***.***>:

…

I was wondering that if the variance or mode are not available in the distribution class, we can also infer them from the predicted samples. Can you add this maybe. — Reply to this email directly, view it on GitHub[#21 (comment)], or unsubscribe[https://github.com/notifications/unsubscribe-auth/AAZQNIJLGMWT5KXW7RHN4ULXOAI7FANCNFSM6AAAAAAZXOQ3HM]. You are receiving this because you were mentioned.[Tracking image][https://github.com/notifications/beacon/AAZQNIKJCMPO2CG4HLXSGZLXOAI7FA5CNFSM6AAAAAAZXOQ3HOWGG33NNVSW45C7OR4XAZNMJFZXG5LFINXW23LFNZ2KUY3PNVWWK3TUL5UWJTTAKBXKO.gif]

[StatMixedML]

Sure. It might make sense to make that behavior explicit so the user knows what they're getting?

Yes, please make sure to adjust the function description to make it clear and self-contained.

[StatMixedML]

Thank you for implementing the changes!

I am currently in the process of completing the unit tests in the upcoming weeks. Once they are finalized, you can modify them accordingly to include a unit test for your specific changes. We can then merge the PR.

[StatMixedML]

can you please also add unit-tests for your function, thanks.