rface.train: Builds an RF-ACE predictor object.
In sivarajankumar/rf-ace: Random Forests with Artificial Contrast Ensembles

Description Usage Arguments Author(s) References See Also Examples

View source: R/rface.train.R

Builds an RF-ACE predictor object.

1	predictorObj <- rface.train(trainData, target, featureWeights = vector(length=0), nTrees = 100, mTry = 10, nodeSize = 3, nMaxLeaves = 1000, quantiles = vector(length=0), nThreads = 1)

`trainData`	A data.frame storing the training data. trainData[featureName] stores a vector of strings (categorical feature) or floats (numerical feature).
`target`	An integer or string, pointing to a feature in trainData, i.e. trainData[target].
`featureWeights`	A vector of nonnegative weights for the features; affects the sampling distribution. By default all features get weight 1 corresponding uniform sampling.
`nTrees`	Number of trees in the forest. Default 100.
`mTry`	Number of randomly sampled candidate features per split. Default 10.
`nodeSize`	Minimum number of train samples per node. Default 3.
`nMaxLeaves`	Maximum number of leaves per tree. Default 1000.
`quantiles`	A vector of quantile points to provide predictions for. If empty, mean prediction will be calculated. Quantiles are only applicable in regression.
`nThreads`	Number of CPU threads to train the model with. Default 1.

Timo Erkkila

http://code.google.com/p/rf-ace

read.afm, rface.filter, rface.predict, rface.save, rface.load

  afmFile <- "test_103by300_mixed_nan_matrix.afm";
  target <- "N:output";
  
  nTrees <- 100;
  mTry <- 30;
  
  nThreads <- 4;

  trainData <- read.afm(afmFile);

  predictorObj <- rface.train(trainData, target, nTrees = nTrees, mTry = mTry, nThreads = nThreads);