ABSTRACT
The conceptualization of things, or objects, as belonging to classes is at the core of all knowledge. We must caution, however, that classes are in the mind of the beholder and that there are often many possible attributes and ways to classify a set of objects. Commonly agreed upon classifications for many objects used routinely are a part of human cultures. We examine here some modern nonlinear methods for classifying and recognizing objects that include clustering, improved clustering, a newer type of fuzzy clustering, probabilistic and fuzzy neural networks, radial basis function neural networks, radial basis functional link nets, ellipsoidal basis function neural networks, ellipsoidal basis functional link nets, and fuzzy ellipsoidal classifiers.
