HIGH-SPEED AND HIGH-ACCURACY LEARNING USING A SELF-ORGANIZATION NETWORK AND A SUPERVISED NETWORK

This paper introduces a feedforward network with a new learning algori thm into which both a self-organized mechanism and a supervised traini ng mechanism are implemented. In particular, from some experiments, it is seen that this network can realize sophisticated signal processing and analysis data with high-speed training ability. The network is di vided into two layers, that is, a self-organized layer and a supervise d layer. However, in the self-organized layer, the training data of th e network also are used for realizing high precision. In addition, the supervised layer also is trained in order to realize quite high accur acy for data recognition. In other words, the training data are used f or both layers. Accordingly, the total precision of the network is imp roved greatly. In addition, the high-speed organization and the parall el processing which are embedded in the original self-organization als o are realized in the new network. In this paper, some comparisons in which conventional neural networks, i.e., multilayer perceptron and se lf-organization models, are used also are shown. In particular, as a n ew application of speech ARMA modelling, a new spectrum envelope estim ation is explained by using this network.