Hierarchical RandomForest


[Up] [Top]

Documentation for package ‘HieRanFor’ version 1.0

Help Pages

HieRanFor-package HieRanFor: A package for running a hierarchical randomForest analysis.
GetMultPropVotes For each case, the multiplicative proportion of votes
HieFMeasure Hierarchical precision, hierarchical recall and hierarchical F measure
ImportanceHie The importance value of each explanatory variable in each local classifier
JoinLevels Combines the factor levels of two input vectors
OliveOilHie Modification of the 'oliveoil' data-set from package 'pdfCluster'.
PerformanceFlatRF Runs a flat classification on the same data as 'hie.RF' and assess performance.
PerformanceHRF Flat and hierarchical performance measures.
PerformanceNewHRF Predict and assess performance of 'new.data', for which the 'true' class is known.
plot.HRF Plot the Hierarchical class structure
PlotImportanceHie Plotting function for the hierarchical importance
predict.HRF For all cases, the proportion of OOB votes that each class received in each local randomForest classifier.
PredictNewHRF Predict crisp class for 'new.data'
RandomHRF Create random class hierarchy with random training and 'new.data' cases.
RunHRF Run the Hierarchical randomForest on the training data.
tuneRF2 the tuneRF function of randomForest after correcting for error relating to errorOld=0