GENETIC ALGORITHMS FOR GENERATION OF CLASS BOUNDARIES

A method is described for finding decision boundaries, approximated by piecewise linear segments, for classifying patterns in R-N,N greater than or equal to 2, using an elitist model of genetic algorithms. It i nvolves generation and placement of a set of hyperplanes (represented by strings) in the feature space that yields minimum misclassification . A scheme for the automatic deletion of redundant hyperplanes is also developed in case the algorithm starts with an initial conservative e stimate of the number of hyperplanes required for modeling the decisio n boundary, The effectiveness of the classification methodology, along with the generalization ability of the decision boundary, is demonstr ated for different parameter values on both artificial data and real l ife data sets having nonlinear/overlapping class boundaries. Results a re compared extensively with those of the Bayes classifier, L-NN rule and multilayer perceptron.