Analytics for the Internet of Things (IoT)

The Gradient Boosting Machines R example

In R, the gbm package can be used to train a GBM. GBMs have the tuning parameters of tree depth and shrinkage. As always, the beauty and danger of R is that it can tune these parameters for you. The following code uses the caret package to fit a GBM model and optimize the tuning parameters:

#make sure needed packages are installed, then load themif(!require(gbm)){ install.packages("gbm")}library(gbm)#reuse train control and training data from random forest example#train the random forest model and specify the number of trees. Use caret to control cross-validation gbmModel <- train (trainData[,predictors], trainData[,target], method = "gbm", trControl = ctrlCV, verbose = FALSE)#run prediction on test ...