A version of the NGE model suitable for fuzzy domains

Abstract

The Nested Generalized Exemplar (NGE) model is an incremental form of inductive learning that generalizes a given training set into hypotheses represented as a set of hyperrectangles in an n-dimensional Euclidean space. The NGE algorithm can be considered a descendent of either Nearest Neighbor (NN) or K-Nearest Neighbor (KNN) algorithms. NGE based systems classify new instances by calculating their similarity to the nearest generalized exemplar (i.e. hyperrectangle). Similarity in an NGE model is implemented by a distance metric namely the Euclidean distance. This paper describes a version of the NGE model suitable for fuzzy domains called Fuzzy NGE (F-NGE). F-NGE learns fuzzy rules for classifying instances into crisp classes. An implementation of F-NGE has been tested in several different knowledge domains for which results are presented and discussed. Results of fuzzy versions of NN and KNN using the same domains are also presented, for comparison.