search

fabsig / GPBoost

Combining tree-boosting with Gaussian process and mixed effects models
Other
574 stars 46 forks source link

Trainig GPBoost with both gp_appox and grouped random effects #146

Closed Shuwen-Liu closed 3 months ago

Shuwen-Liu commented 3 months ago

Dear GPBoost team,

I'm working with large datasets and thus need gp_appox to speed up. However, I can't set grouped random effects at the same time.

This is really important for working with real world dataset. I hope it can be considered as a new enhanced feature of GPBoost.

fabsig commented 3 months ago

This is currently not supported. gp_approx only works if you only have Gaussian processes (via gp_coords) and not grouped random effects.