search

dmlc / XGBoost.jl

XGBoost Julia Package
Other
288 stars 110 forks source link

How to tune hyperparameters of xgboost model ? #100

Open dmahasen opened 2 years ago

dmahasen commented 2 years ago

Hi,

Is there are a way to get metrics values for a given set of parameters in nfold_cv() function? or is there any other way to tune hyperparameters of xgboost model in XGBoost.jl.

Thank you,

Mahasen Dehideniya

ablaom commented 2 years ago

XGBoost has an MLJ interface, which opens up a number of hyper-parameter optimisation strategies. There's even an end-to-end XGBoost/MLJ tutorial on this here. The tutorial does not necessarily represent best practice but does demonstrate the syntax.

I'd probably use RandomSearch in place of Grid.

For an overview of hyper-parameter tuning in MLJ, see this manual entry.