Penerapan Metode Autoregressive Integrated Moving Average (Arima)-Kalman Filter dalam Memprediksi Nilai Anggaran Pendapatan dan Belanja Daerah (Apbd) Provinsi Sumatera Utara Tahun 2023

Abstract

One method that can be combined with the ARIMA forecasting method is the Kalman Filter method, which is a method that can minimize errors by estimating the parameter values. This study aims to predict the value of the North Sumatra Provincial Revenue and Expenditure Budget (APBD) in 2023 by applying the Kalman Filter method to the best ARIMA model, where the data used is APBD value data for 2002-2022. In determining the best ARIMA model, we must first obtain data that is stationary, both in terms of variance and mean. It is from the stationary data that several temporary ARIMA models are identified through the ACF and PACF plots. Then these models were tested with significance, normality, and white noise tests and had a small MAPE value. Next, a simulation of the prediction of the APBD value is carried out and the resulting error value is in the range of 30% -50%. To minimize this error, the Kalman Filter method is applied using degree 1 and 2 polynomials. The combination of these two methods produces an error value of <10%, and the prediction results for the 2023 APBD value are obtained, namely the Regional Revenue value is Rp. 13.52 trillion or an increase of 12.57% and Regional Expenditure is Rp. 15.02 trillion or an increase of 18.82%. From this study it is known that the combination of these two methods produces a relatively small error.