SLAVE: SLAVE model building

Description

This is the internal function that implements the structural learning algorithm on vague environment (SLAVE). It is used to handle classification tasks. Users do not need to call it directly, but just use frbs.learn and predict.

Usage

SLAVE(data.train, persen_cross = 0.6,
    persen_mutant = 0.3, max.iter = 30, max.gen = 30,
    num.labels, range.data.input, k.lower = 0.25,
    k.upper = 0.75, epsilon = 0.1)

Arguments

data.train

a matrix(m x n) of data for the training process, where m is the number of instances and n is the number of variables; The last column is the output variable.

persen_cross

a real number between 0 and 1 representing the probability of crossover.

persen_mutant

a real number between 0 and 1 representing the probability of mutation.

max.iter

the maximal number of iterations.

max.gen

the maximal number of generations for the genetic algorithm.

num.labels

a number of the fuzzy terms.

range.data.input

a matrix containing the ranges of the normalized input data.

k.lower

a lower bound of the noise threshold.

k.upper

an upper bound of the noise threshold.

epsilon

a value between 0 and 1 representing the covering factor.

Details

This method is adopted from A. Gonzalez and R. Perez's paper which is applied for classification problems. SLAVE is based on the iterative rule learning approach which means that we get only one fuzzy rule in each execution of the genetic algorithm. In order to eliminate the irrelevant variables in a rule, SLAVE has a structure composed of two parts: the first part is to represent the relevance of variables and the second one is to define values of the parameters. The following steps are conducted in order to obtain fuzzy rules:

Step 1: Use the genetic algorithm process to obtain ONE RULE for the system.
Step 2: Collect the rule into the final set of rules.
Step 3: Check and penalize this rule.
Step 4: If the stopping criteria is satisfied, the system returns the set of rules as solution. Otherwise, back to Step 1.

This method uses binary codes as representation of the population and applies the basic genetic operators, i.e., selection, crossover, and mutation on it. And, the best rule is obtained by calculating the degree of consistency and completeness.

References

A. Gonzalez, R. Perez, "Selection of relevant features in a fuzzy genetic learning algorithm", IEEE Transactions on Systems, Man, and Cybernetics, Part B: Cybernetics, vol. 31, no. 3, pp. 417 - 425 (2001).