Modelling of infant mortality rate in Nigeria using autoregressive moving average and neural network autoregression

บทคัดย่อ

In the last three decades, the infant mortality rate in Nigeria has been improving but with Nigeria set to miss the Sustainable Development Goal (SDG) target on infant mortality, more need to be done. The study made use of data on infant mortality obtained from the World Bank in determining the infant mortality rate in Nigeria. The trend of infant mortality is identified and models were developed to accurately predict the mortality rate in Nigeria. The time series models; Autoregressive Integrated Moving Average (ARIMA), and Neural Network Autoregression (NNAR); are tested and compared. Model evaluation was done using the five (5) criteria of the Mean Squared Error (MSE), the Root Mean Squared Error (RMSE), the Mean Percentage Error (MPE), the Mean Absolute Error (MAE), and the Mean Absolute Percentage Error (MAPE). Though both models produced similar trends and were found to be suitable for prediction, the NNAR was found to be better than the ARIMA model on all five criteria. Analysis indicates a downward trend in the infant mortality rate in Nigeria and the NNAR model predicts a 12% reduction in the infant mortality rate over the next 10 years.