Residual Analysis of Generalized Autoregressive Integrated Moving Average Bilinear Time Series Model

Abstract

In this study, analysis of residuals of generalized autoregressive integrated moving average bilinear time series model was considered. The adequacy of this model was based on testing the estimated residuals for whiteness. Jarque-Bera statistic and squared-residual autocorrelations were used to test the estimated residuals for whiteness. Generalized autoregressive integrated moving average bilinear time series model was fitted using non linear and non stationary series and the residuals were estimated. The independent test on estimated residuals showed that the residuals were independently distributed. The normality test on the estimated residuals also showed that the residuals followed a normal distribution. The tests on estimated residuals for whiteness were satisfied.

Keywords: Normality test; Residuals; Bilinear model; Jargue-Bera Statistic; Independent test