autoGBM: autoGBM

In 2econsulting/rAutoML:

Description

auto train GBM using pre-defined strategies

Usage

autoGBM(x, y, train_hex, valid_hex, test_hex, model_path = ".",
  max_runtime_secs = 60 * 60, max_models = 60, init_points = 40,
  n_iter = 20)

Arguments

x	independent variables
y	dependent variable
data_hex	H2ODataFrame

Examples

library(h2o)
library(rAutoML)
model_path <- c("C:/tmp")
h2o.init()
data(churn, package = "rAutoML")
churn_hex <- as.h2o(churn)
split_hex <- h2o.splitFrame(data = churn_hex, ratios = c(0.5,0.3), seed = 1234)
train_hex <- split_hex[[1]]
valid_hex <- split_hex[[2]]
test_hex  <- split_hex[[3]]
y <- "Churn."
x <- setdiff(names(churn_hex),  c(y))
autoGBM(x, y, train_hex, valid_hex, test_hex, model_path)

2econsulting/rAutoML documentation built on May 30, 2019, 3:07 a.m.