The “h2o.gbm” Function in R

• Package: h2o

• Purpose: Performs Gradient Boosting Machine (GBM) modeling for classification and regression tasks.

• General class: Modeling

• Required argument(s):

• x: A vector containing the names or indices of the predictor variables.

• y: The name or index of the response variable.

• training_frame: The H2OFrame object containing the training data.

• Notable optional arguments:

• validation_frame: The H2OFrame object containing the validation data.

• nfolds: Specifies the number of folds for cross-validation.

• distribution: The distribution family for the response. Defaults to “AUTO” which chooses the appropriate distribution based on the response data.

• ntrees: The number of trees in the GBM.

• Example:

• # Load the h2o library
  library(h2o)
  
  # Initialize h2o
  h2o.init()
  
  # Import data
  data <- h2o.importFile("https://s3.amazonaws.com/h2o-public-test-data/smalldata/iris/iris_wheader.csv")
  
  # Define predictor and response variables
  predictors <- c("sepal_len", "sepal_wid", "petal_len", "petal_wid")
  response <- "class"
  
  # Split the data into train and test sets
  split <- h2o.splitFrame(data, ratios = 0.8, seed = 123)
  train <- h2o.assign(split[[1]], "train")
  test <- h2o.assign(split[[2]], "test")
  
  # Train the GBM model
  gbm_model <- h2o.gbm(x = predictors, y = response, training_frame = train, validation_frame = test, ntrees = 50)
  
  # View the model summary
  summary(gbm_model)

• This example demonstrates how to train a GBM model using the h2o.gbm function from the h2o package. The function takes predictor variables (x), response variable (y), and training data (training_frame). Optional arguments like validation_frame and ntrees can also be specified. Finally, the summary function is used to view the model summary.