Feature_subset_selection_using_a_genetic_algorithm英文学习资料.pdfVIP

Feature Subset Selection Using A Genetic Algorithm Jiho on Yang and Vasant Honavar  Articial Intelligence Research Group Department of Computer Science Atanaso Hall Iowa State University Ames IA USA fyanghonav ar gcs iastate edu Abstract Many practical pattern classication applicatio ns require a careful selection of attributes or features from a much larger set to represent the patterns to b e classied This feature subset selection problem is a multi criterion optimization problem We prop ose a solution to this problem using a genetic algorithm Our exp eri ments demonstrate the feasibility of this approach for feature subset selection in the automated design of neural network pattern classiers Intro duction Many practical pattern classication tasks eg medical diagnosis require learning of an appropriate classication function that assigns a given input pattern typically represented using a vector of attribute or feature values to one of a nite set of classes The choice of features attributes or measurements used to represent patterns that are presented to a classier aect among other things  The accuracy of the classication function that can b e learned using an inductive learning algorithm eg a decision tree induction algorithm or a neural network learning algorithm The attributes used to describ e the patterns implicitly dene a pattern language If the language is not expressive enough it would fail to capt