Estimation and Prediction for Subset Bilinear Time Series Models with Applications
It has been pointed out earlier that some of the coefficients for the full bilinear models of the form (5.8.1) when fitted to a realisation may be “small” when compared to other coefficients. Therefore, it is useful to see whether it is possible to fit a subset bilinear model to the data which leads to a parsimonious representation. In this chapter, we consider the estimation of a subset bilinear model and we give an algorithm for the estimation of its parameters (see also Gabr and Subba Rao, 1981). The method is illustrated with real data. A comparison is then made between the forecasts obtained from the subset bilinear models and other time series models. Some comments about the transformation of the series are included. (See Subba Rao and Gabr, 1981).
KeywordsSunspot Number Time Series Model Maximum Order Bilinear Model Optimal Forecast
Unable to display preview. Download preview PDF.