Forecasting Harga EUR/USD Harian Menggunakan Long Short-Term Memory (LSTM) Berbasis Log Return
DOI:
https://doi.org/10.29407/6z9g4e51Abstract
Penelitian ini mengkaji penerapan jaringan saraf Long Short-Term Memory (LSTM) untuk meramalkan nilai tukar EUR/USD harian pada periode 2015–2024. Model menggunakan harga Open, High, Low, dan Close (OHLC) yang telah ditransformasi secara logaritmik sebagai fitur input, dengan log harga penutupan hari berikutnya sebagai target prediksi. Delapan konfigurasi LSTM dengan variasi jumlah unit (64 dan 128), jumlah layer (1 dan 2), serta dropout rate (0,2 dan 0,3) dievaluasi secara sistematis menggunakan Root Mean Square Error (RMSE) dan Directional Statistic (Dstat). Metrik Dstat didefinisikan dalam konteks trading: prediksi dianggap benar apabila model secara tepat memperkirakan arah pergerakan harga penutupan esok hari relatif terhadap harga penutupan hari ini. Fungsi Huber loss diterapkan untuk memastikan ketahanan model terhadap data pencilan. Eksperimen dilakukan dengan pembagian data train-test 70:30, di mana StandardScaler hanya di-fit pada data training guna mencegah kebocoran data (data leakage). Hasil menunjukkan bahwa konfigurasi terbaik (C2: 64 unit, 1 layer, dropout=0,3) mencapai Test RMSE sebesar 0,006780 dan Test Dstat sebesar 53,04%, sementara seluruh konfigurasi secara konsisten melampaui baseline acak 50%. Analisis mengkonfirmasi kecenderungan underfitting pada konfigurasi berlapis lebih dalam (C4, C8) serta generalisasi yang baik pada model single-layer.
Keywords:
LSTM, EUR/USD, log return, RMSE, Directional Statistic, forex forecasting##plugins.themes.default.displayStats.downloads##
References
[1] D. C. Yıldırım, I. H. Toroslu, dan U. Fiore, “Forecasting directional movement of Forex data
using LSTM with technical and macroeconomic indicators,” Financial Innovation, vol. 7, no.
1, Des 2021, doi: 10.1186/s40854-020-00220-2.
[2] H. Hassyddiqy dan H. Hasdiana, “Analisis Peramalan (Forecasting) Penjualan Dengan
Metode ARIMA (Autoregressive Integrated Moving Average) Pada Huebee Indonesia,”
Data Sciences Indonesia (DSI), vol. 2, no. 2, hlm. 92–100, Jan 2023, doi:
10.47709/dsi.v2i2.2022.
[3] R. Kumar Ray dkk., “MULTI-MARKET FINANCIAL CRISIS PREDICTION: A
MACHINE LEARNING APPROACH USING STOCK, BOND, AND FOREX DATA,” Int.
J. Appl. Math. (Sofia)., vol. 38, no. 8s, 2025, doi: 10.12732/ijam.v38i8s.602.
[4] O. S. Adesina dan L. O. Obokoh, “A hybrid framework of deep learning and traditional time
series models for exchange rate prediction,” Sci. Afr., vol. 29, Sep 2025, doi:
10.1016/j.sciaf.2025.e02818.
[5] M. R. Pahlevi, “Prediksi Harga Forex Menggunakan Algoritma Long Short-Term Memory,”
JNANALOKA, hlm. 69–76, Sep 2023, doi: 10.36802/jnanaloka.2022.v3-no2-69-76.
[6] M. R. Pahlevi, K. Kusrini, dan T. Hidayat, “Comparison of LSTM and GRU Models for
Forex Prediction,” sinkron, vol. 8, no. 4, hlm. 2254–2263, Okt 2023, doi:
10.33395/sinkron.v8i4.12709.
[7] N. Nengah dkk., “JSIKTI: Jurnal Sistem Informasi dan Komputer terapan Indonesia Deep
Learning Approach for USD to IDR Forecasting with LSTM,” JSIKTI: Jurnal Sistem
Informasi dan Komputer Terapan Indonesia, vol. 8, no. 1, hlm. 91–100, 2025, doi:
10.33173/jsikti.210.
[8] M. Ayitey Junior, P. Appiahene, dan O. Appiah, “Forex market forecasting with two-layer
stacked Long Short-Term Memory neural network (LSTM) and correlation analysis,” Journal of Electrical Systems and Information Technology, vol. 9, no. 1, Des 2022, doi:
10.1186/s43067-022-00054-1.
[9] W. Sri Lestari dan M. Ulina, “Forex Price Predictions using Hybrid TCN-LSTM and LSTMTCN Models,” Indonesian Journal of Artificial Intelligence and Data Mining (IJAIDM), vol.
8, no. 3, hlm. 697–705, 2025, doi: 10.24014/ijaidm.v8i3.38354.
[10]S. Ghahremani dan U. T. Nguyen, “Prediction of foreign currency exchange rates using an
attention-based long short-term memory network,” Machine Learning with Applications, vol.
20, hlm. 100648, Jun 2025, doi: 10.1016/j.mlwa.2025.100648.
[11]W. C. Utomo, “Jurnal Teknologi dan Manajemen Informatika Prediksi Pergerakan Saham
BBRI ditengah Issue Ancaman Resesi 2023 dengan Pendekatan Machine Learning,” vol. 9,
no. 1, hlm. 20–27, 2023, doi: 10.26905/jtmi.v9i1.9135.
[12]A. Saghafi, M. Bagherian, dan F. Shokoohi, “Forecasting Forex EUR/USD Closing Prices
Using a Dual-Input Deep Learning Model with Technical and Fundamental Indicators,”
Mathematics, vol. 13, no. 9, Mei 2025, doi: 10.3390/math13091472.
[13]D. J. Kim, D. H. Kim, dan S. Y. Choi, “Modeling Stylized Facts in FX Markets with
FINGAN-BiLSTM: A Deep Learning Approach to Financial Time Series,” Entropy, vol. 27,
no. 6, Jun 2025, doi: 10.3390/e27060635.
Downloads
Published
Issue
Section
License
This work is licensed under a Creative Commons Attribution-ShareAlike 4.0 International License.
Authors who publish with this journal agree to the following terms:
- Copyright on any article is retained by the author(s).
- The author grants the journal, right of first publication with the work simultaneously licensed under a Creative Commons Attribution License that allows others to share the work with an acknowledgment of the work’s authorship and initial publication in this journal.
- Authors are able to enter into separate, additional contractual arrangements for the non-exclusive distribution of the journal’s published version of the work (e.g., post it to an institutional repository or publish it in a book), with an acknowledgment of its initial publication in this journal.
- Authors are permitted and encouraged to post their work online (e.g., in institutional repositories or on their website) prior to and during the submission process, as it can lead to productive exchanges, as well as earlier and greater citation of published work.
- The article and any associated published material is distributed under the Creative Commons Attribution-ShareAlike 4.0 International License
