Forecasting Indonesian Oil, Non-Oil and Gas Import Export with Fuzzy Time Series
Syalam Ali Wira Dinata(1*), Ayuning Arum Purbosari(2), Primadina Hasanah(3)
(1) Statistics study program, Faculty of Mathematics and Information Technology, Institut Teknologi Kalimantan, Balikpapan, East Kalimantan
(2) Mathematics study program, Faculty of Mathematics and Information Technology, Institut Teknologi Kalimantan, Balikpapan, East Kalimantan
(3) Actuarial Science study program, Faculty of Mathematics and Information Technology, Institut Teknologi Kalimantan, Balikpapan, East Kalimantan
(*) Corresponding Author
Abstract
Indonesia is active in export and import activities. Some of the commodities traded are oil and gas, as well as food and other industrial materials. Export and import activities play a role in determining the stability of the country's economy seen from its trade balance. According to the Central Statistics Agency, Indonesia experienced a deficit of USD 864 million due to a decline in exports at the beginning of 2020. Based on the state of the trade balance, the government needs to make policies in order to maintain the stability of the Indonesian economy. The right decision-making must be supported by accurate information, therefore, through this research, the value of Indonesia's exports and imports will be forecasted in the oil and gas and non-oil and gas sectors for the next period using the Fuzzy Time Series (FTS). FTS was chosen as the forecasting method because it is able to predict free real time data with arbitrary patterns. The data used is data on the value of exports and imports of oil and gas and non-oil and gas sectors for 2011-2020. To overcome the problem of stationary data variance and reduce the error value, a Box Cox transformation will be applied. The research stages include data transformation with Box Cox, forming universe and linguistic sets, determining interval length, fuzzification, forming FLR and FLR, defuzzification and forecasting. The final forecast results estimate that exports and imports in the oil and gas sector in 2021 will decline, while for the non-oil and gas sector will fluctuate and increase from the previous year. Forecasting with Box Cox transform data is more accurate with MAPE 19.56% and RMSE 121.52 compared to forecasting with original data with MAPE 74.89% and RMSE 132.09.
Keywords
Full Text:
PDFReferences
[1] Sutedi, A. (2014). Import Export Law. Achieve Success.
[2] Chen, SM (1996). Forecasting enrollments based on fuzzy time series. Fuzzy Sets and Systems , 81 (3), 311–319.
[3] Winarno, WW (2015). Econometrics and statistical analysis with EViews. Edition 4. UPP STIM YKPN.
[4] Rizal, J., & Akbar, S. (2015). Comparison of Timeseries Data Stationary Tests Between Methods: Control Chart, Correlogram, Dickey Fuller Unit Root, and Degrees of Integration - PDF Free Download. Journal of Gradients, 11 (1), 1040–1046.
[5] Box, GEP, & Cox, DR (1964). An Analysis of Transformations. Journal of the Royal Statistical Society. Series B , 26 (2), 211–252
[6] Lee, MH, Sadaei, HJ, & Suhartono. (2013). Improving TAIEX forecasting using fuzzy time series with Box-Cox power transformation. Journal of Applied Statistics, 40 (11), 2407–2422.
[7] Chang, PC, Wang, YW, & Liu, CH (2007). The development of a weighted evolving fuzzy neural network for PCB sales forecasting. Expert Systems with Applications, 32 (1), 86–96.
[8] Hyndman, Rob. J., & Athanasopoulos, G. (2018). Forecasting: principles and practice . OTEXTS
[9] Spyros Makridakis, & Michele Hibon. (1997). ARMA Models and the Box-Jenkins Methodology. Journal of Forecasting , 16 , 147–163
[10] Johnson, R. A. 2007. Applied Multivariate Statistical Analysis. Vol. 6th Ed. Upper Saddle River, NJ : Pearson/Prentice Hall.
DOI: https://doi.org/10.22146/ijccs.78399
Article Metrics
Abstract views : 1755 | views : 1080Refbacks
- There are currently no refbacks.
Copyright (c) 2022 IJCCS (Indonesian Journal of Computing and Cybernetics Systems)
This work is licensed under a Creative Commons Attribution-ShareAlike 4.0 International License.
View My Stats1