Comparison of highway revenue forecasts using the ARIMA ARIMAX Model

Abstract

This article aims to find the best model by comparing gold jewelry trading forecasts with ARIMA and ARIMAX models from January 2017 to December 2022. for 72 months the precursor variables of ARIMA forecasting were motorway revenue data, dependent variables were motorway revenue forecasting results by ARIMA model and forecasting by ARIMAX model. These are the results of trading forecasts with the ARIMAX model using MAPE, MAE, MSE, RMSE, Ljung-Box, Q (18) and BIC criteria. These criteria were selected from the lowest values ​​and p-values ​​greater than or equal to 0.05. The results show that forecasting highway revenues using the ARIMA (2,1,0) model is more accurate than the ARIMA (2,1,0) model. ARIMAX (3,1,0) The results show that the ARIMA (2,1,0) model for expressway revenue forecasting is more accurate than the ARIMA (2,1,0) model. ARIMAX (3,1,0), with a MAPE value of 0.661 percent, a MAE value of 0.123, and a RMSE value of 0.171.