Sains Malaysiana 48(8)(2019): 1787–1798

http://dx.doi.org/10.17576/jsm-2019-4808-26

Deep Neural Network for Forecasting Inflow and Outflow in Indonesia

(Rangkaian Saraf Dalam untuk Ramalan Aliran Masuk dan Aliran Keluar di Indonesia)

SUHARTONO¹, DIMAS EWIN ASHARI¹, DEDY DWI PRASTYO¹, HERI KUSWANTO¹ & MUHAMMAD HISYAM LEE²*

¹Institut Teknologi Sepuluh Nopember, Surabaya-60111, Indonesia

²Universiti Teknologi Malaysia, 81310 Johor Bahru, Johor Darul Takzim, Malaysia

Received: 1 September 2018/Accepted: 29 May 2019

ABSTRACT

An optimal planning in the preparation of Money Requirement Plan (MRP) by Bank Indonesia is highly beneficial to maintain the availability of money in the community. One of the main factors needed in preparing of MRP is an accurate information about inflow and outflow. This study is to apply Deep Neural Network (DNN) for forecasting inflow and outflow in Indonesia and to compare its performance to ARIMAX as a simpler method and hybrid Singular Spectrum Analysis and DNN (SSA-DNN) as a more complex method. This study focuses on determining the best inputs in DNN, particularly for forecasting time series. A simulation study is used for evaluating the performance of each method related to the patterns in the time series. The real data are monthly inflow and outflow on 5 banknotes denominations from January 2003 to December 2016. The performance was evaluated based on Root Mean Square Error Prediction and Symmetry Mean Absolute Percentage Error Prediction criteria. The results of the simulation study showed that DNN yielded a more accurate forecast than ARIMAX and hybrid SSA-DNN in predicting time series with a trend, seasonal, calendar variation, and nonlinear noise patterns. Moreover, the results of inflow and outflow forecasting showed that DNN provided a more accurate prediction on most all banknotes denominations compared to ARIMAX and hybrid SSA-DNN. In general, these results show that DNN as machine learning model outperforms both ARIMAX as a simpler statistical model and hybrid SSA-DNN as a more complex model.

Keywords: ARIMAX; DNN; inflow and outflow; SSA-DNN; time series forecasting

ABSTRAK

Suatu perancangan yang optimum dalam penyediaan Pelan Wang Keperluan (MRP) oleh Bank Indonesia sangat berfaedah untuk mengekalkan kewujudan wang dalam masyarakat. Salah satu faktor utama yang diperlukan dalam menyediakan MRP adalah maklumat yang tepat tentang aliran masuk dan aliran keluar. Kajian ini bertujuan untuk menerapkan Rangkaian Neural Dalam (DNN) untuk ramalan aliran masuk dan aliran keluar di Indonesia dan untuk membandingkan prestasi ARIMAX sebagai kaedah yang mudah dan hibrid Analisis Spektrum Singular dan DNN (SSA-DNN) sebagai satu kaedah yang lebih kompleks. Kajian ini tertumpu kepada menentukan input terbaik DNN, terutamanya bagi peramalan siri masa. Kajian simulasi yang digunakan untuk menilai prestasi setiap kaedah yang berkaitan dengan corak dalam siri masa. Data sebenar adalah aliran masuk dan aliran keluar bulanan pada 5 denominasi wang kertas dari Januari 2003 untuk Disember 2016. Prestasi dinilai berdasarkan ramalan punca min ralat kuasa dua dan ramalan simetri min mutlak peratusan ralat. Keputusan bagi kajian simulasi menunjukkan bahawa DNN menghasilkan ramalan yang lebih tepat berbanding ARIMAX dan hibrid SSA-DNN untuk meramalkan siri masa dengan kecenderungan, bermusim, perubahan kalendar dan corak bunyi tak linear. Selain itu, keputusan ramalan aliran masuk dan aliran keluar menunjukkan bahawa DNN membuat ramalan yang lebih tepat bagi kebanyakan denominasi wang kertas berbanding ARIMAX dan hibrid SSA-DNN. Secara amnya, keputusan ini menunjukkan bahawa DNN sebagai model pembelajaran mesin yang lebih baik berbanding ARIMAX sebagai model statistik mudah dan hibrid SSA-DNN sebagai model yang lebih kompleks.

Kata kunci: Aliran masuk dan aliran keluar; ARIMAX; DNN; SSA-DNN; waktu ramalan siri

REFERENCES

Abdullah, M.I. & Nor-Muhammad, N.A. 2018. Prediction of colorectal cancer driver genes from Patients' Genome Data. Sains Malaysiana 47(12): 3095-3105.

Ahmad, I.S., Setiawan, S. & Masun, N.H. 2015. Forecasting of monthly inflow and outflow currency using time series regression and ARIMAX: The Idul Fitri Effect. AIP Conference Proceedings 2015: 050002.

Alzahrani, A., Shamsi, P., Dagli, C. & Ferdowsi, M. 2017. Solar irradiance forecasting using deep neural network. Procedia Computer Science 114: 304-313.

Bai, Y., Chen, Z., Xie, J. & Li, C. 2016. Daily reservoir inflow forecasting using multiscale deep feature learning with hybrid models. Journal of Hydrology 532: 193-206.

Bank Indonesia. 2011. Money Circulation. Gerai Info. pp. 1-8.

Bank Indonesia. 2018. Februari 17. Metadata SSKI. http://www.bi.go.id/id/statistik/metadata/SSKI/ Documents/12Metadata%20Uang%20Kartal%20yang%20 Diedarkan.pdf.

Bowerman, B.L. & O'Connell, R.T. 1993. Forecasting and Time Series. Belmont: Wadsworth Publishing Company.

Busseti, E. 2012. Deep Learning for Time Series Modelling. Stanford: Stanford University.

Chen, Z. & Yang, Y. 2004. Assessing forecast accuracy measures. Preprint Series pp. 1-26.

Crone, S.F. & Kourentzes, N. 2009. Input-variable specification for neural network - An analysis of forecasting low and high time series frequency. International Joint Conference on Neural Network doi: 10.1109/IJCNN.2009.5179046.

Golyandina, N. & Zhigljavsky, A. 2013. Singular Spectrum Analysis for Time Series. Heidelberg: Springer-Verlag Berlin.

Goodfellow, I., Bengio, Y. & Courville, A. 2016. Deep Learning. Cambridge: MIT Press.

Gorr, W.L. 1994. Research prospective on neural network forecasting. International Journal of Forecasting 10(1): 1-4.

He, W. 2017. Load forecasting via deep neural networks. Procedia Computer Science 122: 308-314.

Kumar, U. & Jain, V.K. 2010. Time series models (Grey-Markov, Grey Model with rolling mechanism and singular spectrum analysis) to forecast energy consumption in India. Energy 35: 1709-1716.

Lago, J., Ridder, F.D. & Schutter, B.D. 2018. Forecasting spot electricity prices: Deep learning approach and empirical comparison of traditional algorithms. Applied Energy 221: 386-405.

Liu, H., Mi, X. & Li, Y. 2018. Smart multi-step deep learning model for wind speed forecasting based on variational mode decomposition, singular spectrum analysis, LSTM network and ELM. Energy Conversion and Management 159: 54-64.

Makridakis, S. & Hibon, M. 2000. The M3 competition: Results, conclusions, and inmplications. International Journal of Forecasting 16: 451-476.

Makridakis, S., Spiliotis, E. & Assimakopoulos, V. 2018a. Statistical and machine learning forecasting methods: Concerns and ways forward. PLoS ONE 13(3): e0194889.

Makridakis, S., Spiliotis, E. & Assimakopoulos, V. 2018b. The M4 competition: Results, findings, conclusions and way forward. International Journal of Forecasting. DOI: 10.1016/j.ijforecast.2018.06.001.

Otok, B.W., Suhartono, Ulama, B.S.S. & Endharta, A.J. 2011. Design of experiment to optimize the architecture of wavelet neural network for forecasting the tourist arrivals in Indonesia. Communications in Computer and Information Science 253(3): 14-23.

Setiawan, Suhartono, Ahmad, I.S. & Rahmawati, N.I. 2015. Configuring calendar variation based on time series regression method for forecasting of monthly currency inflow and outflow in Central Java. AIP Conference Proceedings 1691 2015: 050024.

Suhartono, Amalia, F.F., Saputri, P.D., Rahayu, S.P. & Ulama, B.S.S. 2018a. Simulation study for determining the best architecture of multilayer perceptron for forecasting nonlinear seasonal time series. Journal of Physics: Conference Series 1028(1): 012214.

Suhartono, Saputri, P.D., Prastyo, D.D. & Rahayu, S.P. 2018b. Hybrid quantile regression neural network model for forecasting currency inflow and outflow in Indonesia. Journal of Physics: Conference Series 1028(1): 012213.

Suhartono, Setyowati, E., Salehah, N.A., Lee, M.H., Rahayu, S.P. & Ulama, B.S.S. 2017. A hybrid singular spectrum analysis and neural networks for forecasting inflow and outflow currency of Bank Indonesia. In. Soft Computing in Data Science, edited by Yap B., Mohamed A. & Berry, M. Singapore: Springer Nature. pp. 3-18.

Suhartono. 2007. Feedforward neural network for time series forecasting. Yogyakarta: PhD Dissertation, Gadjah Mada University (Unpublished).

Vautard, R. & Ghil, M. 1989. Singular spectrum analysis in nonlinear dynamics, with applications to paleoclimatic time series. Physica D: Nonlinear Phenomena 35: 395-424.

Yu, C., Li, Y. & Zhang, M. 2017. Comparative study on three new hybrid models using Elman Neural Network and Empirical Mode Decomposition based technologies improved by Singular Spectrum Analysis for hour-ahead wind speed forecasting. Energy Conversion and Management 147: 76-85.

Zhang, X., Wang, J. & Zhang, K. 2017. Short-term electric load forecasting based on singular spectrum analysis and support vector machine optimized by Cuckoo search algorithm. Electric Power Systems Research 146: 270-285.

Zubaidi, S.L., Dooley, J., Alkhaddar, R.M., Abdellatif, M., Al-Bugharbee, H., Ortega. & Martorell, S.A. 2018. Novel approach for predicting monthly water demand by combining singular spectrum analysis with neural networks. Journal of Hydrology 561: 136-145.

*Corresponding author; email: mhl@utm.my