FORECASTING URBAN POPULATION USING REGRESSION AND TIME SERIES MODELS: A CASE STUDY OF ASTANA

Abstract

This study addresses the problem of urban population prediction using both regression and time series models. A number of linear regression models based on various combinations of socioeconomic and demographic variables were created and evaluated, along with an ARIMA (AutoRegressiveIntegrated Moving Average) model for time series prediction. The models were assessed using key performance metrics including R2, AIC, BIC, MAE,RMSE, and MAPE. Labor force statistics combined with gross regional product data generated highly precise and comprehensible prediction models according to comparative analysis. These findings create useful information that helps urban planners execute their work and policymakers build effective population dynamics programs.