ABSTRACT
From this, we obtained the model of the process in terms of the three trained neural networks that were connected according to Fig. 18. According to this figure, given some input data set, the network NNmem identifies the degree of membership of a sample in each of the two classes. Network NNi predicts productivity for class "well cleaned", and NN2 for class "poorly cleaned". Input information is fed to all three networks. The outputs of all three networks are then fed into special elements which process the networks outputs to determine the weighted sum and as a result predict final process effectiveness and the degree of cleanliness.
