Extended Kalman Filter In Recurrent Neural Network: USDIDR Forecasting Case Study


Muhammad Asaduddin Hazazi(1), Agus Sihabuddin(2*)

(1) Master Program of Computer Science and Electronics, FMIPA UGM, Yogyakarta
(2) Department of Computer Science and Electronics, Universitas Gadjah Mada
(*) Corresponding Author


Artificial Neural Networks (ANN) especially Recurrent Neural Network (RNN) have been widely used to predict currency exchange rates. The learning algorithm that is commonly used in ANN is Stochastic Gradient Descent (SGD). One of the advantages of SGD is that the computational time needed is relatively short. But SGD also has weaknesses, including SGD requiring several hyperparameters such as the regularization parameter. Besides that SGD relatively requires a lot of epoch to reach convergence. Extended Kalman Filter (EKF) as a learning algorithm on RNN is used to replace SGD with the hope of a better level of accuracy and convergence rate. This study uses IDR / USD exchange rate data from 31 August 2015 to 29 August 2018 with 70% data as training data and 30% data as test data. This research shows that RNN-EKF produces better convergent speeds and better accuracy compared to RNN-SGD.


exchange rates; forecasting; recurrent neural network; stochastic gradient descent; extended Kalman Filter

Full Text:



[1] A. Hector, M. Claudio dan S. Rodrigo, 2002, Artificial Neural Networks in Time series Forecasting: A Comparative Analysis, Kybernetika, Vol. 38, No. 6, pp. 685-707.

[2] S.E. Rumagit, 2012, Prediksi Pemakaian Listrik dengan Menggunakan Jaringan Syaraf Tiruan dan ARIMA di Wilayah Suluttenggo, Tesis, Program Studi S2 Ilmu Komputer, Universitas Gadjah Mada, Yogyakarta.

[3] E. Munarsih, 2011, Penerapan Model ARIMA - Neural Network Hybrid Untuk Peramalan Data Time Series, Tesis, Program Studi S2 Matematika, Universitas Gadjah Mada, Yogyakarta.

[4] D.U. Wutsqa, R. Kusumawati dan R. Subekti, 2014, The Application of Elman Recurrent Neural Network Model for Forecasting Consumer Price Index of Education, Recreation, and Sports in Yogyakarta, IEEE. 10th International Conference on Natural Computation, 192-196.

[5] A. Nandury dan L. Sherry, 2016, Anomaly Detection in Aircraft Data Using Recurrent Neural Network (RNN), Integrated Communications,Navigation, and Surveillance (ICNS), Herndon, VA, pp. 5C2-1-5C2-8.

[6] L. Susanti, A. Fariza, Setiawardhana, 2011, Peramalan Harga Saham Menggunakan Recurrent Neural Network dengan Algoritma Backpropagation Through Time (BPTT), Skripsi,Politeknik Elektronika Negeri Surabaya, Surabaya.

[7] J.A. Perez-Ortiz, J. Calera-Rubio dan M.L. Forcada, 2001, Online Text Prediction with Recurrent Neural Network, Neural Processing Letter 12: 127-140, Kluwer Academic Publisher, Netherlands.

[8] F. Syahrian, 2016, Perbandingan Metode Optimasi Stochastic Gradient Descent, Adadelta, Dan Adam Pada Jaringan Saraf Tiruan Dalam Klasifikasi Data Aritmia, Tesis, Program Studi S2 Ilmu Komputer, Universitas Gadjah Mada, Yogyakarta.

[9] A.S. Prabowo, A. Sihabuddin, A.S. Nugrahito, 2019, Adaptive Moment Estimation On Deep Belief NetworkFor Rupiah Currency Forecasting, Indonesian Journal of Computing and Cybernetics Systems (IJCCS), Vol.13, No.1 pp. 31~42, Yogyakarta.

[10] W. Xu, 2011, Towards Optimal One Pass Large Scale Learning with Averaged Stochastic Gradient Descent, arxiv, July 13

[11] M. Cernansky dan L. Benuskova, 2003, Simple Recurrent Neural Network Trained By RTRL and Extended Kalman Filter Algorithm, Neural Network World, 13(3), pp. 223-234.

[12] P. Trebaticky, 2005, Recurrent Neural Network Training with Extended Kalman Filter, M.Bielikova (Ed.), IIT.SRC 2005, April 27, 2005, pp. 57-64.

[13] R. Adnan, F.A. Ruslan, A.M. Samad dan Z.M. Zain, 2013, New Artificial Neural Network and Extended Kalman Filter Hybrid Model of Flood Prediction System, IEEE 9th International Colloquium on Signal Processing and its Applications, 8 - 10 Maret 2013, Kuala Lumpur, Malaysia.

[14] A.N. Cernodub, 2014, Training Neural Network for Classification Using The Extended Kalman Filter: A Comparative Study, SSN 1060 992X, Optical Memory and Neural Networks (Information Optics), Vol. 23, No. 2, pp. 96- 103, Allerton Press, Inc.

DOI: https://doi.org/10.22146/ijccs.47802

Article Metrics

Abstract views : 3893 | views : 3300


  • There are currently no refbacks.

Copyright (c) 2019 IJCCS (Indonesian Journal of Computing and Cybernetics Systems)

Creative Commons License
This work is licensed under a Creative Commons Attribution-ShareAlike 4.0 International License.

Copyright of :
IJCCS (Indonesian Journal of Computing and Cybernetics Systems)
ISSN 1978-1520 (print); ISSN 2460-7258 (online)
is a scientific journal the results of Computing
and Cybernetics Systems
A publication of IndoCEISS.
Gedung S1 Ruang 416 FMIPA UGM, Sekip Utara, Yogyakarta 55281
Fax: +62274 555133
email:ijccs.mipa@ugm.ac.id | http://jurnal.ugm.ac.id/ijccs

View My Stats1
View My Stats2