Comparative Performance of Simple Exponential Smoothing, Brown’s Linear Trend and ARIMA Model on Forecasting Neonatal Mortality Rate in Nigeria
Christogonus Ifeanyichukwu Ugoh *
Department of Statistics, Faculty of Physical Sciences, Nnamdi Azikiwe University, Awka, Nigeria.
Nneka Chidinma Nwabueze
Department of Statistics, Faculty of Physical Sciences, Nnamdi Azikiwe University, Awka, Nigeria.
Nwabueze Achunam Simeon
Department of Community Medicine, Faculty of Medicine, Nnamdi Azikiwe University, Nnewi, Nigeria.
Eze Theophine Chinaza
Department of Statistics, Faculty of Physical Sciences, Nnamdi Azikiwe University, Awka, Nigeria.
Okafor Chinasa Ogedi
Department of Paediatrics, Faculty of Paediatrics Medicine, COOUTH, Nigeria.
*Author to whom correspondence should be addressed.
Abstract
Paper proposes an appropriate time series model that is used to forecast the NMR in Nigeria. The data used for the study is sourced from the World Bank for a period of 1980-2019. The ARIMA model and Exponential Smoothing are fitted on the raw data. The Bayesian Information Criterion (BIC) is adopted to assess the adequacy of the ARIMA models. The NMR series is stationary after the second differencing. The ARIMA (0,2,0) with BIC value of -3.358 is considered the appropriate model among other ARIMA models, and it is compared to SES and Brown’s LT using Theil’s U Statistics and MAPE. The results showed that the Brown’s LT model is more ideal and adequate for forecasting NMR in Nigeria based on the Theil’s U forecast accuracy measures of 0.001911, and that by 2030, Nigeria will have a reduced NMR of 31.5 deaths per 1,000 live births, which shows a drop to 21.5%.
Keywords: NMR, exponential smoothing, BIC, ARIMA, SES, Brown’s LT, Theil’s U statistic