The attached benchmark code includes a 3-fold cross-validation step and the usual training-prediction procedure. The 2nd step cost me less than a minute on an 8-core CPU.
To install the R-package of xgboost, please first install the latest version of devtools, then run
devtools::install_github('dmlc/xgboost',subdir='R-package')
Windows users may need to install RTools to compile. A detailed guide is here:https://github.com/dmlc/xgboost/tree/master/demo/kaggle-otto
XGBoost also offers python interface. For more information of XGBoost, please visit https://github.com/dmlc/xgboost.
1 Attachment —