COUNTERPROPAGATION NEURAL-NETWORK FOR MODELING A CONTINUOUS CORRELATED PROCESS

Variations of a counterpropagation (CP) neural network model were inve stigated for developing both a predictive model and an inverse mapping of process manufacturing operations. A brief explanation of the manuf acturing process is provided, along with a review of recent neural net work applications to manufacturing. Variations of Kohonen's self organ izing map are discussed with respect to the counterpropagation algorit hm. Methods are presented for determining the underlying distribution in a data set.