renll / rongba

Source code for the course project "RoNGBa: A Robustly Optimized Natural Gradient Boosting Training Approach with Leaf Number Clipping"

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

RoNGBa: A Robustly Optimized Natural Gradient Boosting Training Approach with Leaf Number Clipping

This is a course project rushed in a weekend with the outcome of the preprint. The code is mostly the same as the original ngboost repository, except for minor changes on the hyperparameters and the learner settings.

Simply run the following command:

bash run.sh

to reproduce the results reported in our preprint.

About

Source code for the course project "RoNGBa: A Robustly Optimized Natural Gradient Boosting Training Approach with Leaf Number Clipping"

License:Apache License 2.0


Languages

Language:Python 86.6%Language:Shell 13.2%Language:Makefile 0.3%