A Comparative Statistical Analysis of Simple and Double Exponential Smoothing Models: An Application to Libya's Population Data (1960- 2024)

Abstract

This study aims to apply the Exponential Smoothing method to forecast the population of Libya during the period 1960–2024, with a comparison between the Simple and Double Exponential Smoothing models to identify the most accurate and suitable model for predicting population time series. The study adopts a descriptive-analytical approach, where population data were collected and analyzed to identify the general trends of the time series, and then both models were implemented using Minitab software.

The results indicate that Libya’s population has followed an overall increasing trend in recent years. The comparison between the models shows a clear difference in forecasting accuracy, with the Double Exponential Smoothing model achieving the lowest values for all accuracy measures (MSD, MAD, MAPE), making it the most suitable for future population forecasts. Predictions using the double model indicate a continuous increase in population during the period 2025–2034, reflecting the ongoing upward trend.

The study recommends using the Double Exponential Smoothing model as the primary forecasting tool for population, with regular updates as new data become available, and utilizing the results for planning in education, health, and housing to meet future population needs.