GiniImportanceTree {rfVarImpOOB}

computes Gini information gain for one tree from randomForest

Description

computes importance scores for an individual tree.

These can be based on Gini impurity or Accuracy or logloss

Usage

GiniImportanceTree(bag, RF, k, ylabel = "Survived", returnTree = FALSE, 


    zeroLeaf = TRUE, score = c("PMDI21", "MDI", "MDA", "MIA")[1], 


    Predictor = Mode, verbose = 0)

Arguments

bag

data to compute the Gini gain for

RF

object returned by call to randomForest()

k

which tree

ylabel

name of dependent variable

returnTree

if TRUE returns the tree data frame otherwise the aggregated Gini importance grouped by split variables

zeroLeaf

if TRUE discard the information gain due to splits resulting in n=1

score

scoring method:PMDI=mean decrease penalized Gini impurity (note:the last digit is the exponent of the penalty!),

MDI=mean decrease impurity (Gini), MDA=mean decrease accuracy (permutation),

MIA=mean increase accuracy

Predictor

function to estimate node prediction, such as Mode or mean or median. Alternatively, pass an array of numbers as replacement for the yHat column of tree

verbose

level of verbosity

Value

if returnTree==TRUE returns the tree data frame otherwise the aggregated Gini importance grouped by split variables

Author(s)

Markus Loecher <[e-mail]>

Examples

rfTit = rfTitanic(nRows = 500,nodesize=10)


rfTit$data$Survived = as.numeric(rfTit$data$Survived)-1


k=1


tmp <- InOutBags(rfTit$RF, rfTit$data, k)


IndivTree =getTree(rfTit$RF,k)


#plot(as.party(tmp))#does not work


InTree = GiniImportanceTree(tmp$inbag,rfTit$RF,k,returnTree=TRUE)


OutTree = GiniImportanceTree(tmp$outbag,rfTit$RF,k,returnTree=TRUE)