Enhance plot_pdp and fix plot_scatter_submodels

[AlexAndorra]

• Add env YML files, especially for dev
• Add softmax_link option to plot_pdp for the Categorical likelihood case
• Fix plot_scatter_submodels to be able to handle the categorical likelihood case

[AlexAndorra]

Thanks for the ping amigo! I didn't forget, I'm just swamped right now, but will push the change ASAP El El vie, 7 mar 2025 a la(s) 3:40, Osvaldo A Martin < ***@***.***> escribió:

…

***@***.**** commented on this pull request. ------------------------------ In pymc_bart/utils.py <#218 (comment)>: > @@ -396,11 +396,18 @@ def identity(x): p_d = _sample_posterior( all_trees, X=fake_X, rng=rng, size=samples, excluded=excluded, shape=shape ) + # need to apply func to full array and to last dimension if it's softmax + if func.__name__ == "softmax": @AlexAndorra <https://github.com/AlexAndorra> — Reply to this email directly, view it on GitHub <#218 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AHIJMTG3TLOPJJZINE3BT2D2TFLODAVCNFSM6AAAAABXICSTZ2VHI2DSMVQWIX3LMV43YUDVNRWFEZLROVSXG5CSMV3GSZLXHMZDMNRWGU4TMNJZGQ> . You are receiving this because you were mentioned.Message ID: ***@***.***>

[aloctavodia]

I just added the change to plot_ice myself. You already helped a lot with the other changes. Thanks @AlexAndorra.