Real-time pattern recognition. I. Neural network algorithms for normal models

Traditional neural network models, when contrasted with normal models, have some distinct similarities and differences. The paper describes the salient features of perceptrons and their extensions, relative to normal models. Conjunctoid neural network models when contrasted with normal models also have some relative advantages, the paper describes these advantages and disadvantages. Normal neural network models are introduced which combine some of the features of the conjunctoid and traditional normal models. The paper also introduces the algorithm to implement a normal neural network model for the airplane example described in an earlier part of the paper. The paper briefly describes the architectures for implementing the algorithms on parallel machines, along with conclusions and future research directions.<<ETX>>