aml_gbm: AML Gradient Boosted Machine (GBM)

In jmwerner/ArtisanalMachineLearning: Hand Made Pure Organic Machine Learning

Description

Trains an ensemble of trees via gradient boosting

Usage

aml_gbm(data, response, learning_rate = 0.1, n_trees = 10, m = NULL,
  evaluation_criterion = sum_of_squares, min_obs = 5, max_depth = 8,
  verbose = FALSE)

Arguments

data	Input data.frame of dimension n x p for training the GBM
response	Response vector of size nx1 corresponding to the training data
learning_rate	Shrinkage factor used to dictate learning speed, defaults to .1
n_trees	Number of trees to train, defaults to 10
m	Number of columns to randomly use at each splitting iteration, defaults to all columns
evaluation_criterion	Function that calculates error criterion for fitting, defaults to sum of squares
min_obs	Minimum observations allowed to end up in a single node, defaults to 5
max_depth	Maximum number of successive splits allowed to happen in the tree, defaults to 8
verbose	Flag to display training updates in the console

Value

Results trained list of class aml_random_forest filled with random forest trees

jmwerner/ArtisanalMachineLearning documentation built on Jan. 7, 2020, 1:50 a.m.