Timeline for How to tune hyperparameters of xgboost trees?
Current License: CC BY-SA 3.0
9 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| May 19, 2016 at 20:33 | comment | added | geneorama |
For the unbalanced class issue, scale_pos_weight is now documented in the parameter documentation. scale_pos_weight is not a caret tuning parameter, but you can compare manually. In my case, using the weight happened to have little effect (binary classification, >20% positives)
|
|
| Feb 16, 2016 at 7:52 | comment | added | tchakravarty | @lovedynasty You should probably ask a new question with exactly what you are looking for. | |
| Feb 16, 2016 at 7:31 | comment | added | discipulus |
What would be the changes required for multiclass classification. Also documentation says use scale_pose_weight for imbalanced classification. Can you provide details on how to? Thanks!
|
|
| Dec 14, 2015 at 7:13 | comment | added | phiver | @ML_Pro, at the moment these paramaters: nrounds, max_depth, eta, gamma, colsample_bytree, min_child_weight. You could always create your own model and expand the possible parameters for the grid search. | |
| Nov 13, 2015 at 14:37 | comment | added | GeorgeOfTheRF | That is support by xgboost right? My question is about which all parameters does caret support for grid search | |
| Nov 13, 2015 at 14:36 | vote | accept | GeorgeOfTheRF | ||
| Jun 10, 2017 at 18:10 | |||||
| Nov 13, 2015 at 14:03 | comment | added | tchakravarty |
@ML_Pro Support for most xgboost parameters now exists, in particular support for gamma is new. Here is a full list of supported parameters.
|
|
| Nov 13, 2015 at 13:56 | comment | added | GeorgeOfTheRF | Does caret still only support eta, gamma and max depth for grid search what about subsample and other parameters of xgboost? | |
| Nov 13, 2015 at 13:33 | history | answered | tchakravarty | CC BY-SA 3.0 |