cvboost: Gradient boosting for regression and classification with...

In xnie/rlearner: R-learner for Heterogeneous Treatment Effect Estimation

Description

Gradient boosting for regression and classification with cross validation to search for hyper-parameters (implemented with xgboost)

Usage

cvboost(
  x,
  y,
  weights = NULL,
  k_folds = NULL,
  objective = c("reg:squarederror", "binary:logistic"),
  ntrees_max = 1000,
  num_search_rounds = 10,
  print_every_n = 100,
  early_stopping_rounds = 10,
  nthread = NULL,
  verbose = FALSE
)

Arguments

x	the input features
y	the observed response (real valued)
weights	weights for input if doing weighted regression/classification. If set to NULL, no weights are used
k_folds	number of folds used in cross validation
objective	choose from either "reg:squarederror" for regression or "binary:logistic" for logistic regression
ntrees_max	the maximum number of trees to grow for xgboost
num_search_rounds	the number of random sampling of hyperparameter combinations for cross validating on xgboost trees
print_every_n	the number of iterations (in each iteration, a tree is grown) by which the code prints out information
early_stopping_rounds	the number of rounds the test error stops decreasing by which the cross validation in finding the optimal number of trees stops
nthread	the number of threads to use. The default is NULL, which uses all available threads. Note that this does not apply to using bayesian optimization to search for hyperparameters.
verbose	boolean; whether to print statistic

Value

a cvboost object

Examples

## Not run: 
n = 100; p = 10

x = matrix(rnorm(n*p), n, p)
y = pmax(x[,1], 0) + x[,2] + pmin(x[,3], 0) + rnorm(n)

fit = cvboost(x, y, objective="reg:squarederror")
est = predict(fit, x)

## End(Not run)

xnie/rlearner documentation built on April 11, 2021, 12:49 a.m.