You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
I’ve been testing the model on my regression dataset , which includes several multimodal and long-tailed continuous features. While CTAB-GAN+ is expected to handle such features well, the generated data diverges significantly from the real distributions, especially visible in KDE and ECDF plots.
My question is:
Are there additional configurations needed to better capture multimodal and skewed distributions in regression tasks?