LkbGBMLMichigan {FKBL}R Documentation

Creates a knowledge base

Description

This is the implementation of Michigan genetic method. This is a genetic algorithm where each individual represents a rule. With the given probability, the crossing operation is performed by swapping two partition labels between two individuals. With the given probability, the mutation is performed by with a 50% of probability changing to a random appropriate value each of the partitions labels. "label" means a reference to a division in the partition, coded as a positive integer. This two operations are done the number of iterations specified. This makes the population grow and diversify. The final result is a knowledge base made by gathering in a base the actual pupulation of rules. Described in chapter 5, pages 105-117 at Ishibuchi et al.\

Usage

 LkbGBMLMichigan(P, gen=1000, cross=0.9, muta=0.01,  train)

Arguments

Takes the vector of partitions, the number of generations, the crossing probability, the mutation probability and the train data.

P The vector of partitions.
gen The number of generations.
cross The cross probability up to 1.
muta The mutation probability up to 1.
train The train dataset.

Value

Returns a knowledge base with the partitions and the rules.

Source

begin{itemize}

  • Ishibuchi, H., Nakashima, T., Nii, M.
  • "Classification and modeling with linguistic information granules."
  • Soft Computing Approaches to Linguistic Data Mining.
  • Springer-Verlag, 2003 end{itemize}

    Examples

     data(P)
     data(trainA)
     LkbGBMLMichigan(P,1000,0.9,0.01,trainA)
    

    [Package FKBL version 0.50-4 Index]