simpleRF: simpleRF

In mnwright/simpleRF: Random Forests with emphasis on simplicity

simpleRF

Description

Implements Random Forests (Breiman 2001) with emphasis on simplicity. Uses reference classes and only plain R. Not optimized for computation speed. Allows rapid prototyping of RF-type algorithms.

Usage

simpleRF(formula, data, num_trees = 50, mtry = NULL,
  min_node_size = NULL, replace = TRUE, probability = FALSE,
  splitrule = NULL, unordered_factors = "ignore", num_threads = 1)

Arguments

formula	Object of class formula or character describing the model to fit.
data	Training data of class data.frame.
num_trees	Number of trees.
mtry	Number of variables to possibly split at in each node.
min_node_size	Minimal node size. Default 1 for classification, 5 for regression, 3 for survival and 10 for probability estimation.
replace	Sample with replacement. Default TRUE.
probability	Grow a probability forest. Default FALSE.
splitrule	Splitrule to use in trees. Default "Gini" for classification and probability forests, "Variance" for regression forests and "Logrank" for survival forests.
unordered_factors	How to handle unordered factor variables. One of "ignore", "order_once", "order_split" and "partition" with default "ignore".
num_threads	Number of threads used for mclapply, set to 1 for debugging.

Details

Unordered factor variables can be handled in different ways. Use "ignore" to treat them as ordered in the order of the factor levels. With "order_once" and "order_split" they are ordered by their response values. For "order_once" this is done once before the analysis, for "order_split" this is done in each split. With "partition" all 2-partitions of the factor levels are considered for splitting.

Author(s)

Marvin N. Wright

References

Breiman, L. (2001). Random forests. Mach Learn, 45(1), 5-32.

Examples

library(simpleRF)

# Classification
simpleRF(Species ~ ., iris)

# Prediction
train_idx <- sample(nrow(iris), 2/3 * nrow(iris))
iris_train <- iris[train_idx, ]
iris_test <- iris[-train_idx, ]
rf_iris <- simpleRF(Species ~ ., data = iris_train)
pred_iris <- rf_iris$predict(iris_test)
table(iris_test$Species, pred_iris)

mnwright/simpleRF documentation built on March 28, 2022, 8:57 a.m.