Comparison of neural networks to statistical techniques forprediction of time series generated by nonlinear dynamic systems

摘要

The following paper is focused on comparison of neural networks tostatistical techniques for time series prediction. Four statisticalmodels, the ARIMA, the exponential smoothing, the exponential growth andthe bilinear model are compared to two neural network architectures, thehierarchical multilayer perceptron and the ontogenic cascade correlationnetwork. The intercomparison was done on two examples, a generic and areal-world one. The results of analyses were most promising from theneural networks point of view