Abstract:
This research aimed to compare of combined forecasting methods for volume of natural gasoline exported. The data used in this study is a monthly time series secondary dataset collected from the Department of Energy Policy and Planning Office Ministry of Energy, covering the period from January 2016 to December 2022, total of 84 months, were divided into two sets. Set 1 includes data from January 2016 to December 2021, total of 72 months, used to create forecasting models. By using 3 individual forecasting methods including Box-Jenkins method, Winters Additive Exponential Smoothing Method, Time Series Regression Method and 4 Combined Forecasting Method including the Equivalent Weighted. Method (EW), the Inverse of Mean Square Error Method (IMSE), Combined Forecast Method using Regression Analysis (RG) and Unequivalent Weighted Method (UNEW). Set 2 consists of data from January to December 2022, total of 12 months, used for model comparison. The researchers compared the accuracy of models using the Mean Absolute Percentage Error (MAPE) and found that the most suitable forecasting method is the combined forecasting method using regression analysis, because it has the lowest MAPE value