Evaluation for Prediction Accuracies of Parallel-type Neuron Network

The parallel-type neuron network (PNN) is researched to improve on the decrease in capabilities of the neuron network by the interference of the learning caused between the outputs of BP network (BPN) of two outputs or more and the difficulty of the common achievement of the middle layer used for each output. The research to compare prediction accuracies of nonlinear time series signals prediction systems using BPN and PNN has been performed so far. However, it has not attained demonstrating the existence of dominance of all prediction accuracies of PNN to BPN. Then, the experimental evaluation of the dominance of all outputs of PNN which could exist for the theory by results of the comparison of learning rules of BPN and PNN was performed using nonlinear time series signals prediction systems in this research. As a result, the dominance was showed.