Forecasting the Value of Indonesia’s Exports using Model Hybrid Arimax-NN
Article Sidebar
Main Article Content
Abstract
Time sequence data often exhibits both linear and nonlinear patterns, which can lead to inaccurate forecasts when using traditional methods that are limited to capturing only one type of pattern. To address this limitation, this study employs a hybrid method that combines the strengths of Autoregressive Integrated Moving Average with Exogenous Variables (ARIMAX) and neural networks (NN). The ARIMAX model effectively captures linear patterns, while the NN excels at modeling nonlinearities. The primary objective of this research is to optimize the ARIMAX-NN hybrid model for forecasting Indonesia’s export values. Through rigorous model selection, the ARIMAX ([1,5,12],1,0)-NN 1 neuron model emerged as the best-performing configuration, achieving the lowest Mean Absolute Percentage Error (MAPE), Root Mean Squared Error (RMSE), and Mean Absolute Deviation (MAD) values. The forecasts for January December 2024 reveal a pattern of decreasing export values during the month of Eid al-Fitr, a trend consistent with historical patterns and economic insights.
Article Details
This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.
References
Siswanti, T. E., Yanti, T. S. Pemodelan ARIMAX (Autoregressive ntegrated Moving Average with Exogenous Variable). Statistics Proceedings. 2020; 6(2):113-8.
Andreas, C., Sediono, Ana, E., Suliyanto, Fadillah, M., Mardianto. Application of the ARIMAX-GARCH Model in Modeling and Forecasting Electronic Money Transaction Volume in Indonesia. Journal of Mathematics Education, Science and Technology. 2021; 6(2):241-56.
Silvia, R. H., Achmad, A. I. Application of the ARIMAX Method with the Effect of Calendar Variation on the Price Forecasting of Cayenne Pepper Commodity in West Java Province. Statistics. 2023; 3(2):689-98.
Kusumaningrum, N., Purnamasari, I., Siringoringo, M. Forecasting Using the ARIMAX-NN Hybrid Model for Total Non-Cash Payment Transactions. Journal of Statistics and Its Application on Teaching and Research. 2023; 5(1):1-14.
Sawitri, M. N. D., Sumarjaya, I. W., Tastrawati, N. K T. Forecasting using the Backpropagation Neural Network Method. Journal of Mathematics, 2018; 7(3):264-70.
Hyndman, R. J., Athanasopoulos, G. (2018). Forecasting: principles and practice. OTexts.
Eksiandayani, S., Suhartono, Prastyo, D.D. Hybrid ARIMAX-NN Model for Forecasting Inflation. Proceeding International Conference on Science, Technology and Humanity. 2015;181-187.
Ramadhan, R.W., Iqbal, F., Utamy, N. P. Ananda, A. N. The Influence of Oil and Gas and Non-Oil and Gas Sector Exports on Indonesia’s GDP. Journal of Management and Social Economics. 2023; 6(2):62-71.
Ministry of Home Affairs. Three Consecutive Years of Surplus, Indonesia’s Trade Balance in April 2023 Exceeded USD3.94 Billion. Available from: Three Consecutive Years of Surplus, Indonesia’s Trade Balance in April 2023 Exceeded USD 3.94 Billion - Ministry of Trade of the Republic of Indonesia (kemendag.go.id).Regional Science 1992;32:467-86.
Puspitaningrum, D. Introduction to Artificial Neural Networks. Yogyakarta: ANDI; 2006.
Rachman, A. S., Chollisodin, I., and Fauzi, M. A. Forecasting of Sugar Production Using the Backpropagation Artificial Neural Network Method in PG Candi Baru Sidoarjo. Journal of Information Technology and Computer Science Development. 2018; 2(4):1683-9.
Noon, J. Artificial Neural Networks and Their Programs Use Matlab. Yogyakarta: ANDI Publishers; 2005.
Zhang, G. P. Time series forecasting using a hybrid ARIMA and neural network model. Neurocomputing. 2003; 50:159-75.
Putera, M. L. S. Forecasting Non-Cash Transactions Using ARIMAX-NN with Calendar Configuration. BAREKENG: Journal of Mathematical and Applied Sciences. 2020; 14(1):135-46.
Murni, C. K. Comparison of Beverage Sales Forecasting Using Single Exponential Smoothing and Triple Exponential Smoothing Algorithms. Journal of Informatics Development. 1(2):59-64.
Novianto, Y., Nataliani, Y. Rainfall Forecasting by Moon Grouping Using Brown’s Double Exponential Smoothing Method. Journal of Information Systems and Technology. 2022; 10(4):347-354.