tyler-tomita / RandomerForest

Discriminant Projection Forest results, datasets, etc.
44 stars 21 forks source link

Learn the art of tuning XGBoost #104

Closed tyler-tomita closed 7 years ago

tyler-tomita commented 7 years ago

Since this is the state-of-the-art tree ensemble, we'd like to compare RerF to this.

Which parameters is it most sensitive to?

What are reasonable ranges of parameter values to try in order to balance computational cost and classification performance?

Try running initially on a subset of Trunk and sparse parity datasets