Bilinear Time Series Model for Estimating a Disease Death Rate

We compare new time-series methods for estimating the death rate of an emerging and re-emerging disease and our approach is based on One-Dimensional Integrated Autoregressive Bilinear Time Series Model and Generalized Integrated Autoregressive Bilinear Time Series Model. The parameters of the proposed models are estimated using Newton-Raphson iterative method and statistical properties of the derived estimates are investigated. An algorithm was proposed for fitting the two models. To determine the best order of the models, Akaike Information Criterion (AIC) was adopted. Residual variance was used to see which model perform better. We illustrated the new concept with real life data. One-Dimensional Integrated Autoregressive Bilinear Time Series Model outperforms Generalized Integrated Autoregressive Bilinear Time Series Model in the estimation of death rate of a disease