Simon Dedman
Simon Dedman
mate please could you send me your run script and your data (or a representative chunk so it'll run)? Thanks.
gbm.auto: report around L1036 populated by `Bin_Bars$var`, from L858: `summary(get(Bin_Best_Model)` from L686 can be bin best simp if worthy. So 858 should populate 1036 with simp thus make simplified bars,...
I just tried the first run with only tc 1, lr 0.01 and bf 0.5. Best combo was the unsimplified version, so even though Report.csv lists the simp predictors dropped...
gaus: L1143 & 4: ``` Report[1:(length(Gaus_Bars[,1])),(reportcolno - 2)]
Pushed change, model re-run by Frances didn't need simplifying so change not tested, dangerzone.
NFF any update on this, did the change solve the issue? If so please mark as closed. Cheers!
Emailed Paul Metcalfe asking why gbm.fit isn't backwards compatible with gbm.step 2018.10.11, main function params look identical Emails Jane asking again about her/RObert's updated gbm.predict.grids code which is apparently faster.
Discussed with Brandon @ gbm Feb 2019, recent update to gbm was bugfixes, state of gbm3 unknown, could just try it out.
Could overhaul gbm with xgboost - see https://xgboost.readthedocs.io/en/latest/R-package/xgboostPresentation.html & related docs, can restructure gbm.auto to use this instead? Has inbuilt parallel & potentially better performance. Could also interface with CUDA...
Use CatBoost instead, is better than all others. https://github.com/catboost/catboost https://catboost.ai/en/docs/installation/r-installation-binary-installation https://catboost.ai/en/docs/concepts/r-quickstart