XGB Vignettes

Click Here for three vignettes on XGB cross validation and hyperparameter tuning

I have uploaded three coding vignettes which show how to fit xgboost models in python. The focus of the videos is proper hyperparameter tuning using early stopping, which requires deviation from standard ‘sklearn’ practices (maybe there is a better way but I don’t know of one). A very good article on this which is mirrored somewhat closely by our vignettes is Hyperparameter Tuning XGBoost with early stopping by Jeff Macaluso.

Click below to watch the videos:

• XGB Vignette 1: Simple CV with Early Stopping
• XGB Vignette 2: Grid Search CV with Early Stopping
• XGB Vignette 3: Random Search CV with Early Stopping

The code for the vignette can be

• xgboost_vignette or
• xgboost_vignette_live.ipynb

And if you want to really follow along, the dataset can be downloaded here:

• UCI Credit Card Default