Bayesian Probabilistic Projection of Population Census in the Kingdom of Saudi Arabia
Saheed A. Afolabi *
Department of Mathematics, College of Computing and Mathematics, King Fahd University of Petroleum and Minerals, Dhahran 31261, Kingdom of Saudi Arabia.
*Author to whom correspondence should be addressed.
Abstract
Population census supplies a complete and accurate picture of a country's population and its residents' characteristics. Modeling population growth has been worked upon by different scholars before now with more of a classical approach and less of a Bayesian approach. Therefore, an attempt is made in this work to apply Bayesian probabilistic projection on the usual exponential growth rate model in estimating population parameters and predicting population census in the Kingdom of Saudi Arabia (KSA) across thirteen (13) regions. The obtained data from WorldData and United Nations Population were used for the estimation and projection with the application of appropriate prior, likelihood, and posterior selection through Bayesian inference. This approach is reasonably accurate and well-calibrated with a significant precision of 0.01025 approximately 99% model accuracy for the period due to the estimated population parameters that were used: to make a comparison with the 2019 Population Census of Saudi Arabia which was perfectly closed and to forecast for the next 80 years using out-sample cases.
Keywords: Population growth rate, conjugate prior, Predictive Inference (PI), Highest Density Interval (HDI), projection