Open PyDataBlog opened 4 years ago
Thanks for bringing it ot my attention. Gonna have a good read
@PyDataBlog @xiaodaigh This is an awesome idea! There is a whole new generation of these models the predict an entire conditional distribution. Julia would be an awesome environment for this. See my discussion: https://discourse.julialang.org/t/mlj-probabilisitc-models-basics/36009/3?u=albert_zevelev
A new kid on the gradient boosting block is the NGBoost variant described in this project from the Standford ML Group. It would be really nice to have this implementation in this project as well.