Control for evtree

Share:

Description

Various parameters that control aspects of the evtree fit.

Usage

1
2
3
4
5
evtree.control(minbucket = 7L, minsplit = 20L, maxdepth = 9L, 
  niterations = 10000L, ntrees = 100L, alpha = 1,
  operatorprob = list(pmutatemajor = 0.2, pmutateminor = 0.2,
    pcrossover = 0.2, psplit = 0.2, pprune = 0.2),
  seed = NULL, ...)

Arguments

minbucket

the minimum sum of weights in a terminal node.

minsplit

the minimum sum of weights in a node in order to be considered for splitting.

maxdepth

maximum depth of the tree. Note, that the memory requirements increase by the square of the m'aximum tree depth.

niterations

in case the run does not converge, it terminates after a specified number of iterations defined by niterations.

ntrees

the number of trees in the population.

alpha

regulates the complexity part of the cost function. Increasing values of alpha encourage decreasing tree sizes.

operatorprob

list or vector of probabilities for the selection of variation operators. May also be specified partially in which case the default values are still used for the unspecified arguments. Always scaled to sum to 100 percent.

seed

an numeric seed to initialize the random number generator (for reproducibility). By default the seed is randomly drawn using runif in order to inherit the state of .Random.seed. If set to seed = -1L, the random number generator is initialized by the system time.

...

additional arguments.

Value

A list with the (potentially processed) control parameters.