<![CDATA[The Galur Office of Religious Affairs (KUA) routinely records marriage data but has not optimally utilized it for operational planning. The absence of predictive analysis complicates the allocation of marriage registrars and administrative resources, particularly during fluctuations in marriage applications. This study aims to develop a forecasting model for monthly marriages using the Autoregressive Integrated Moving Average (ARIMA) method to support data-driven decision-making. Historical data covering 120 months (January 2016–December 2025) were divided into 80% training and 20% testing data. The Augmented Dickey-Fuller (ADF) test indicated stationarity at level; therefore, first-order differencing was not required but was retained as a candidate specification. Based on residual diagnostics and the lowest AIC/BIC values, ARIMA(0,1,1) was selected as the best model. Testing produced a Mean Absolute Error (MAE) of 6.84 and a Root Mean Squared Error (RMSE) of 8.37. The final model forecasts a constant 13 marriages per month throughout 2026. Although its point-forecast accuracy did not consistently outperform a simple baseline, ARIMA provides a statistical inference framework that can support proactive staffing and administrative resource planning at KUA Galur.]]>
Copyrights © 2026
