RandomForestClassifier: Random Forest Classifier

In Mthrun/Classifiers: Common Supervised Machine Learning Algorithms

Random Forest Classifier

Description

Training of a random forest classification, if required also predicts classification on test data

Usage

RandomForestClassifier(TrainData, TrainCls, TestData, Names,
NumberOfTrees=500, VariableImportance=TRUE, Seed,
PlotIt=TRUE, Verbose=FALSE, ABCanalysis=FALSE, Fast=FALSE)

Arguments

TrainData	(1:n,1:d) matrix, data Array of n cases withd variables of TrainData or Full data
TrainCls	vector, Array of variable names
TestData	Optional, (1:m,1:d) matrix, data Array of d cases with n variables of TestData
Names	Optional, (1:d) vector, Array of variable names, if not given, colnames are used
NumberOfTrees	Number of trees to grow. This should not be set to a too small number, to enshure that every input row gets predicted at least a few times
VariableImportance	Should importance of predictors be assessed?
Seed	set a seed for randomization
PlotIt	wether to plot error versus trees
Verbose	whether to show results of forest, default FALSE
ABCanalysis	Defualt FALSE, If TRUE select indizes of group A of feautures with highest accuracy values by computed ABCanalysis, only if VariableImportance==TRUE
Fast	default FALSE ( randomForest is used. TRUE: randomForestSRC is used but VariableImportance and ABCanalysis is set FALSE, because its not implemented for this case.

Details

reiman's random forest algorithm (based on Breiman and Cutler's original Fortran code) for classification. The “local” (or casewise) variable importance is computed as follows: For classification, it is the increase in percent of times a case is OOB and misclassified when the variable is permuted, in more detail:

the prediction error on the out-of-bag portion of the TrainData is recorded (error rate for each tree). Then the same is done after permuting each predictor variable. The difference between the two are then averaged over all trees, and normalized by the standard deviation of the differences. If the standard deviation of the differences is equal to 0 for a variable, the division is not done (but the average is almost always equal to 0 in that case).

Value

list V with

Classification	[1:n], classification by randomForest of k clusters
ImportancePerVariable	[1:d,1:2], Importance of Variables by Gini and Accuracy
ContingencyTable	Vergleich TrainCls zu Classification
ImportancePerClass	Importance of Variables per Cluster j of k clusters of TrainCls
Forest	Object of randomForest
MostImportantFeatures	Feautures Index of the first m most-important features defined by group A ob ABCanalysis
TestCls	Null if no TestData Given, [1:m] vector of k classes otherwise

Author(s)

Michael Thrun

References

Breiman, L. (2001), Random Forests, Machine Learning 45(1), 5-32.

Breiman, L (2002), “Manual On Setting Up, Using, And Understanding Random Forests V3.1”, https://www.stat.berkeley.edu/~breiman/Using_random_forests_V3.1.pdf.

Examples

library(FCPS)
data("Chainlink")
Data=Chainlink$Data
Cls=Chainlink$Cls
split=Classifiers::splitquoted(Data,Cls,Percentage = 80)
out=RandomForestClassifier(split$TrainData,TrainCls = split$TrainCls,TestData = split$TestData)
table(out$TestCls,split$TestCls)

Mthrun/Classifiers documentation built on June 28, 2023, 9:28 a.m.