Maximizes a Sdt value by monte carlo method

Share:

Description

Takes all cues in fftree and builds blocksize random generated fftrees. If no better fftree is found, or the the value didn't improve greater than threshold, the algorithm finishes. Use Fftm instead of calling this method directly.

Usage

1
2
fftm.montecarlo(tree, whatToMaximize, blocksize, threshold, cluster = NULL,
  cores = 1)

Arguments

tree

fftree

whatToMaximize

stringvector; Which parameters of the tree should be chased? The index of the vector represents the sequence. See Sdt for possible values.

blocksize

integer; How many random trees should be generated, before threshold is checked?

threshold

double; How much of an improvement in whatToMaximize should be found to generate another block?

cluster

If you have a HPC cluster, specifiy configuration here. See makeCluster from snow package.

cores

Integer; If multithreading is wanted (recommended for bigger computations!), you can specify your cores here.

Details

You can chase all values provided by Sdt

Value

Fftree

See Also

Fftm, bruteforcemaximize, chase, maximize

Examples

1
2
rawdata <- fftm.Titanic.data()
tree <- Fftm(Survived ~ Age + Sex + Class, "montecarlo",rawdata, whatToMaximize = "percCorr", blocksize = 1000, threshold=.001, cores=4)

Want to suggest features or report bugs for rdrr.io? Use the GitHub issue tracker.