An International Publisher for Academic and Scientific Journals
Author Login
Scholars Journal of Physics, Mathematics and Statistics | Volume-4 | Issue-02
Lower Triangular Vector Bilinear Autoregressive Time Series Model
Iberedem A. Iwok
Published: May 22, 2017 |
111
73
DOI: 10.36347/sjpms
Pages: 44-52
Downloads
Abstract
This work proposed a strategy for fitting a subset form of a vector bilinear model. In this case, lower triangular
portion of the vector linear and non-linear components of a vector autoregressive bilinear structure was considered. The
workability of the method was illustrated using a 3 – dimensional vector of time series. The models were found to fulfil
the assumptions of model adequacy and each time dependent variable depended only on a subset of lagged time varying
quantities under consideration. Apart from being parsimonious, the fitted models were found to perform better than the
parent vector bilinear autoregressive models as revealed by some statistical comparative analysis.