Comparison of different methods of bilinear time series prediction

Prediction has applications in time series analysis, feedforward control, and many other situations in which knowledge of the future behavior of process is desired. Algorithm of prediction results from a model of process or phenomenon. In this paper minimum variance bilinear predictor for the general form of bilinear time series model is considered. Bilinear predictor is compared with neural network predictor (NNP) and group method of data handling (GMDH) Predictor.