Operator GeneticFeatureSelection

A Feature Selection operator.This operator uses a genetic algorithm to select a subset of TheAttributes. It calls C4.5 to evaluate the individuals of the genetic population. TheTargetAttribute specifies which attribute is the target attribute for the learning algorithm whose performance is used to select the best feature subset. PopDim gives the size of the population for the genetic algorithm. StepNum gives the number of generations. The probabilities of mutation and crossover are specified with ProbMut and ProbCross.

Parameter

Parameter Object Type optional min_arg max_arg Remarks
TheInputConcept Concept Input no 1 1 inherited
TheAttributes BaseAttribute Input no 1  
TheTargetAttribute BaseAttribute Input no 1 1
SampleSize Value Input no 1 1 positive integer
TheOutputConcept Concept Output no 1 1 inherited
PopDim Value Input no 1 1 positive integer; try 30
StepNum Value Input no 1 1 positive integer; try 20
ProbMut Value Input no 1 1 real between 0 and 1; try 0.001
ProbCross Value Input no 1 1 real between 0 and 1; try 0.9