The aim of this study is to obtain an optimal model between the traditional time series model (ARIMA) and Weighted Markov Chain. The historical dataset of U5MR in Nigeria from 1980-2019 is obtained from the official website of World Bank. ARIMA modeling involved differencing of the data to attain stationarity, while WMC involved classification of the datasets into clusters using k-means cluster analysis and transition of states. Two performance measures Theil’s U Statistic and MAPE are used to evaluate the two models based on in-sample and out-sample. The results shows that ARIMA(0,3,2) is a better model to forecast U5MR in Nigeria.
Citation: Ugoh C.I., Osuji G.A., Nwankwo C.H., Nwabueze N.C., Eze T.C., Orji G.O. (2022) Application of Autoregressive Integrated Moving Average Model and Weighted Markov Chain on Forecasting Under-Five Mortality Rates in Nigeria, European Journal of Statistics and Probability, Vol.10, No.2, pp., 39-53,
Keywords: ARIMA, K-Mean Cluster, MAPE, Theil’s U Statistic, U5MR, Weighted Markov Chain (WMC)
This work by European American Journals is licensed under a Creative Commons Attribution-NonCommercial-NoDerivs 3.0 Unported License