makeFeatSelControlExhaustive: Create control structures for feature selection.

Description

The following methods are available:

Usage

makeFeatSelControlExhaustive(same.resampling.instance = TRUE,
    maxit = as.integer(NA), max.features = as.integer(NA))
  makeFeatSelControlGA(same.resampling.instance = TRUE,
    maxit = as.integer(NA), max.features = as.integer(NA),
    comma = FALSE, mu = 10, lambda, crossover.rate = 0.5,
    mutation.rate = 0.05)
  makeFeatSelControlRandom(same.resampling.instance = TRUE,
    maxit = 100L, max.features = as.integer(NA),
    prob = 0.5)
  makeFeatSelControlSequential(same.resampling.instance = TRUE,
    method, alpha = 0.01, beta = 0.01,
    maxit = as.integer(NA), max.features = as.integer(NA))

Arguments

Value

[FeatSelControl]. The specific subclass is one of FeatSelControlExhaustive, FeatSelControlRandom, FeatSelControlSequential, FeatSelControlGA.

Details

[object Object],[object Object],[object Object],[object Object]

The GA is a simple (mu, lambda) or (mu + lambda) algorithm, depending on the comma setting. A comma strategy selects a new population of size mu out of the lambda > mu offspring. A plus strategy uses the joint pool of mu parents and lambda offspring for selecting mu new candidates. Out of those mu features, the new lambda features are generated by randomly choosing pairs of parents. These are crossed over and crossover.rate represents the probability of choosing a feature from the first parent instead of the second parent. The resulting offspring is mutated, i.e. its bits are flipped with probability mutation.rate. If max.features is set, offspring are repeatedly generated until the setting is satisfied.