search

ebecht / infinityFlow

25 stars 8 forks source link

Advice to tune the hyperparameters in extra_args_regression_params in training SVM, NN and XGBoost models #23

Open denvercal1234GitHub opened 1 year ago

denvercal1234GitHub commented 1 year ago

Hi there,

It is a very useful tool, so I aim to really make it work despite all the issues. So, thanks again for your patience and help.

The tutorial (https://bioconductor.org/packages/devel/bioc/vignettes/infinityFlow/inst/doc/training_non_default_regression_models.html) is very useful to run these various models. Although it was not the intention and probably out of the scope of the tutorial, do you have any advice or resources that can help me choose the values for various hyper parameters in extra_args_regression_params for flow data?

Is it sufficient to use what are specified in the vignettes, or these values should be adjusted per dataset? Is the gain if adjusted worth it?

Thanks again.