EUVIMEDEUVIMED entdecken
Open Access · CC BY-SA 4.0

Optimization of Hybrid ARIMAX-GRU Hyperparameters Using Bayesian Optimization for Antam Gold Price Forecasting

Armalia Kusuma Putri, Trimono Trimono, Muhammad Nasrudin

bit-Tech · 10.08.2026 · 9 (1) · 1209-1220

DOI: 10.32877/bt.v9i1.4078 ↗

Kurz gefasst

Fachartikel in der Zeitschrift bit-Tech, online erschienen am 10.08.2026. Verfasst von Armalia Kusuma Putri und 2 weiteren Autor:innen. Der Volltext ist frei zugänglich (Open Access, CC BY-SA 4.0).

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Gold is widely recognized as a safe-haven investment; however, its price remains highly volatile due to macroeconomic factors, making accurate forecasting challenging. This study proposes a Hybrid ARIMAX–GRU framework for one-step-ahead daily forecasting of Antam gold prices using daily data collected from April 2016 to December 2025. The forecasting framework incorporates the Rupiah exchange rate and the Bank Indonesia policy rate as exogenous variables. The ARIMAX model captures the linear component of gold price movements, while the GRU model learns the remaining nonlinear patterns from the ARIMAX residuals. Bayesian Optimization is employed to determine the GRU hyperparameter configuration. Model performance is evaluated on a hold-out testing set using Mean Absolute Error (MAE), Root Mean Squared Error (RMSE), and Mean Absolute Percentage Error (MAPE). The proposed Hybrid ARIMAX–GRU model achieves an RMSE of Rp34,320.75 and a MAPE of 0.6691%, compared with Rp34,692.66 and 0.6734%, respectively, obtained by the standalone ARIMAX model, corresponding to relative improvements of approximately 1.07% in RMSE and 0.64% in MAPE, while producing a slightly higher MAE. Furthermore, the Diebold–Mariano test indicates that the difference in predictive accuracy between the two models is not statistically significant at the 5% significance level. These findings suggest that the proposed hybrid framework provides a modest improvement in forecasting performance for the evaluated testing period, although additional comparative studies are required to isolate the contribution of Bayesian Optimization and to further validate the model under different forecasting scenarios.

Zitieren

APA

Putri, A. K., Trimono, T., & Nasrudin, M. (2026). Optimization of Hybrid ARIMAX-GRU Hyperparameters Using Bayesian Optimization for Antam Gold Price Forecasting. bit-Tech, 9(1), 1209-1220. https://doi.org/10.32877/bt.v9i1.4078

Vancouver

Putri AK, Trimono T, Nasrudin M. Optimization of Hybrid ARIMAX-GRU Hyperparameters Using Bayesian Optimization for Antam Gold Price Forecasting. bit-Tech. 2026;9(1):1209-1220. doi:10.32877/bt.v9i1.4078

BibTeX
@article{Putri2026,
  author = {Putri, Armalia Kusuma and Trimono, Trimono and Nasrudin, Muhammad},
  title = {Optimization of Hybrid ARIMAX-GRU Hyperparameters Using Bayesian Optimization for Antam Gold Price Forecasting},
  journal = {bit-Tech},
  year = {2026},
  volume = {9},
  number = {1},
  pages = {1209--1220},
  doi = {10.32877/bt.v9i1.4078}
}
Kostenlos starten →