statmixedml / catboostlss

An extension of CatBoost to probabilistic modelling

CatBoostLSS - An extension of CatBoost to probabilistic modelling

We propose a new framework of CatBoost that predicts the entire conditional distribution of a univariate response variable. In particular, CatBoostLSS models all moments of a parametric distribution, i.e., mean, location, scale and shape (LSS), instead of the conditional mean only. Choosing from a wide range of continuous, discrete and mixed discrete-continuous distributions, modelling and predicting the entire conditional distribution greatly enhances the flexibility of CatBoost, as it allows to gain additional insight into the data generating process, as well as to create probabilistic forecasts from which prediction intervals and quantiles of interest can be derived. In the following, we provide a short walk-through of the functionality of CatBoostLSS.

News

Repo is currently not maintained.

Reference Paper

catboostlss's Issues

ETA for this code?

Project dead ?

Hi, this project look awesome and very usefull!!!

The python files are all empty, is this project still Alive ?

Thanks!

Question: Catboost Quantile Regression

Would you be able to shed some light on when to use the distribution fitting features of CatBoostLSS versus the quantile regression feature of the standard CatBoost package?

how to install?

Interesting project. hope to experiment with it.

ETA for package?

Anxiously awaiting this package as it looks very useful and interesting. what is timeline for deployment? All files are currently empty, when will something be shared? imperfect as is still better than not available imho.

Is it already usable?

Hi Alexander,
is it already usable? I can't find any code inside - just empty files...

All files are empty

Hello! Could you say why all files are empty? (in python and catboostlss folder)

Any intention of doing this with lightGBM

Just wondering :)

Is this forthcoming?

I know there was some discussion here, in the Catboost and NGBoost repos about this. Any updates?