gbm2sas: Export a Gradient Boosted Regression Model (GBM) to SAS
In az0/mlmeta: Machine Learning Metaprogramming

Description Usage Arguments Details Examples

Generates SAS DATA step code to predict the values of a GBM model.

1	gbm2sas(fit, n.trees = fit$n.trees, drop = TRUE)

`fit`	the model fitted by `gbm`
`n.trees`	the number of trees to export
`drop`	whether to drop the variables for the individual trees

This function supports unordered factors, ordered factors, and missing values. No special considerations are needed for the Gaussian distribution, and for other distributions, the user is responsible for the link function in SAS.

require(gbm)
mtcars.gbm <- gbm(mpg ~ cyl + disp, data = mtcars, bag.fraction = 1.1,
    distribution = "gaussian")
mtcars.sas <- gbm2sas(mtcars.gbm)
cat(mtcars.sas, file='mtcars.sas')