GENERATING-SHRINKING ALGORITHM FOR LEARNING ARBITRARY CLASSIFICATION

This paper proposes a novel generating-shrinking algorithm that builds and then shrinks a three-layer feedforward neural network to achieve arbitrary classification in n-dimensional Euclidean space. The algorit hm offers guaranteed convergence to a 100% correct classification rate on training patterns. Decision regions resulting from the algorithm a re analytically described, so the generalisation behaviour of the trai ned network is analytically known. By altering the value of a referenc e number, the trained neural classifier can achieve scale-invariant ge neralisation as well as equal-distance generalisation to accommodate d ifferent requirements.