XGB Lithology Classification Lessons Learned from Force 2020 ML Competition

David J. Emery, Marcelo Guarido, Daniel O. Trad

The CREWES Data Science Initiative is back with the free webinars and the next learning lab is October 14th, at 4pm (MT). David Emery is the presenter and he will show insights from well log data analysis and lithology classification with XGBoost in Python.

FORCE Machine Predicted Lithology was a competition in 2020 to do lithology classification from well logs. It was a challenging dataset that required heavy pre-processing: data cleaning, data imputation, scaling and normalization, and the target (lithology) also is imbalanced. Different methodologies were required just to get to the point of the machine learning modelling, that was done with the use of the XGBoost classifier algorithm. And all those solutions will be presented by David.

For this event, David will do a hands-on demonstration on how to use a the XGBoost library in Python.