Enhancing Demographic Forecasts: An Application of Adaptive Linear Neuron (Adaline) and ARIMA Models to Iraqi Fertility Rates Data

Volume 1 , Issue 1 , April 2026 , Pages 19-28

Authors

Dr. Mohammed M. Faqe 1 ; Dr. Akhter Khan S. Hamad 2 ; Soran H. Mohamad 2

1 Statistics and Informatics, Administration and Economics – Suleimani, Iraq

2 Statistics and Informatics, Administration and Economics – Sulaimani, Iraq

DOI logo 10.17656/jzsb.12296

Keywords

Abstract

The projection of fertility rates has tremendous implications for how we manage resources, develop policy, and plan demographic groups, especially in countries experiencing social and economic change like Iraq. The purpose of this study was to advance demographic forecasting by analysing the predictive accuracy of two time-series models, namely an Autoregressive Integrated Moving Average (ARIMA) model and an Adaptive Linear Neuron (Adaline) Artificial Neural Network. An analysis of annual fertility rate data from 1950–2025, with stationarity tests and suitable transformations applied before developing the forecasting models, will be conducted. Forecasting accuracy, mean squared error (MSE), and root mean square error (RMSE), will be measured. Although ARIMA (3,1,3) provides a reliable means of capturing linear information, an Adaline-ANN model achieves a much higher level of predictive performance because it can capture complex temporal dynamics. This was shown by the substantial degree to which the neural network produced lower error values than the ARIMA model and by the greater stability with which they provided estimates for the period of time from 2026 to 2035. Overall, these findings support the effectiveness and adaptability of an Adaline-ANN approach to fertility forecasting and emphasize the significance of data-driven learning approaches for providing a more reliable basis for predictions and for assisting in the evidence-based planning of demographics.

Statistics
  • Article view156
  • Downloads16
  • First online30 April 2026
  • Published at30 April 2026

  • RIS
  • BibTeX
  • EndNote
  • Mendeley
  • APA (7th edition)
  • MLA (9th edition)
  • Chicago
  • Harvard
  • IEEE
  • Vancouver