Hyperparameter Optimization for Lightgbm

[brianreinke95]

Hi. This is the only site I found when searching for hyperparameter optimization: https://unit8co.github.io/darts/userguide/hyperparameter_optimization.html#hyperparameter-optimization-with-optuna Those are for NN models. About gridsearch: Each forecasting models in Darts provides a gridsearch() method for basic hyperparameter search. This method is limited to very simple cases, with very few hyperparameters, and working with a single time series only. I'm looking for a way to tune my multi-series lightgbm model. Is there a way? Thanks, Brian.

[madtoinou]

Hi @brianreinke95,

In the documentation, we recommend using other libraries such as Optuna or Ray-tune to optimize hyper-parameters instead of gridsearch().

The tutorial you linked also applies to regression models such as LightGBMModel, you just need to adapt the objective(trial) function to train and evaluate the model. Aside from the callback/checkpointing, there is nothing specific to NN.