Pattern classification using genetic algorithms: determination of H

Abstract

A methodology based on the concept of a variable string length GA (VGA) is developed for determining automatically the number of hyperplanes for modeling the class boundaries in a GA-classifier. The genetic operators and fitness function are defined to take care of the variability in chromosome length. It is proved that the method is able to arrive at the optimal number of misclassifications after a sufficiently large number of iterations, and will need a minimal number of hyperplanes for this purpose. Experimental results on different artificial and real life data sets demonstrate that the classifier, using the concept of a variable length chromosome, can automatically determine an appropriate value of the number of hyperplanes, and also provide performance better than that of the fixed length version. Its comparison with another approach using a VGA is provided.