A Novel Gated Fusion CNN-LSTM Model for Multi-Horizon Intraday Gold Price Forecasting
DOI:
https://doi.org/10.4108/eetiot.12806Keywords:
gold price forecasting, CNN-LSTM, gated fusion, deep learning, time series forecasting, directional accuracyAbstract
Gold serves as a key inflation hedge and portfolio stabilizer, making accurate price forecasting essential for investors. Hourly gold prices exhibit pronounced non-linearity and microstructure noise that limit traditional econometric models, motivating a shift toward deep learning. Existing CNN-LSTM architectures cascade convolutional and recurrent layers sequentially, without a mechanism to reconcile their complementary representations. We propose a Dual-Branch CNN-LSTM architecture with Gated Fusion, combining a convolutional-recurrent deep branch with a parallel raw-input skip branch, adaptively merged by a learned gate inspired by the Gated Multimodal Unit. Input sequences are restructured via sliding windows across three forecast horizons (24→1, 48→2, and 72→3 hours). The model is validated on 37,140 hourly XAUUSDm observations (Exness, 2020-2026) against 1D-CNN, LSTM, and Sequential CNN-LSTM baselines, achieving the best or tied-best accuracy across all configurations. For the 24-hour horizon, MAE = 9.00 USD, R² = 0.9994, and DA = 52.6%, on the original USD scale. Diebold-Mariano tests confirm significant gains over the 1D-CNN and Sequential CNN-LSTM baselines (p < 0.001), while McNemar tests confirm directional-accuracy gains in 8 of 9 comparisons (p < 0.05). These results show a modest but robust improvement over prior CNN-LSTM designs, offering a reliable foundation for risk management pending economic backtesting.
Downloads
References
[1] Zhang C, Sjarif NNA, Ibrahim R. Deep learning models for price forecasting of financial time series: a review of recent advancements: 2020-2022. WIREs Data Min Knowl Discov. 2024;14(1):e1519. doi:10.1002/widm.1519.
[2] Li W, Law KLE. Deep learning models for time series forecasting: a review. IEEE Access. 2024;12:92306-92327. doi:10.1109/ACCESS.2024.3422528.
[3] Foroutan P, Lahmiri S. Deep learning systems for forecasting the prices of crude oil and precious metals. Financ Innov. 2024;10(1):111. doi:10.1186/s40854-024-00637-z.
[4] Livieris IE, Pintelas E, Pintelas P. A CNN-LSTM model for gold price time-series forecasting. Neural Comput Appl. 2020;32(23):17351-17360. doi:10.1007/s00521-020-04867-x.
[5] You W, Chen J, Xie H, Ren Y. Which uncertainty measure better predicts gold prices? New evidence from a CNN-LSTM approach. N Am J Econ Finance. 2025;76:102375. doi:10.1016/j.najef.2025.102375.
[6] Kong X, et al. Deep learning for time series forecasting: a survey. Int J Mach Learn Cybern. 2025;16(7-8):5079-5112. doi:10.1007/s13042-025-02560-w.
[7] Nallamothu S, Rajyalakshmi K, Arumugam P. Gold price prediction using skewness and kurtosis based generalized auto-regressive conditional heteroskedasticity approach with long short term memory network. J Inst Eng India Ser B. 2024;105(6):1715-1727. doi:10.1007/s40031-024-01070-7.
[8] Cohen G, Aiche A. Forecasting gold price using machine learning methodologies. Chaos Solitons Fractals. 2023;175:114079. doi:10.1016/j.chaos.2023.114079.
[9] Jabeur SB, Mefteh-Wali S, Viviani JL. Forecasting gold price with the XGBoost algorithm and SHAP interaction values. Ann Oper Res. 2024;334(1-3):679-699. doi:10.1007/s10479-021-04187-w.
[10] Abu-Doush I, Ahmed B, Awadallah MA, Al-Betar MA, Rababaah AR. Enhancing multilayer perceptron neural network using archive-based Harris hawks optimizer to predict gold prices. J King Saud Univ Comput Inf Sci. 2023;35(5):101557. doi:10.1016/j.jksuci.2023.101557.
[11] Weng F, Chen Y, Wang Z, Hou M, Luo J, Tian Z. Gold price forecasting research based on an improved online extreme learning machine algorithm. J Ambient Intell Humaniz Comput. 2020;11(10):4101-4111. doi:10.1007/s12652-020-01682-z.
[12] Memon BA, Tahir R, Naveed HM, Cheng K. Forecasting gold and platinum prices with an enhanced GRU model using multi-headed attention and skip connection. Miner Econ. 2025. doi:10.1007/s13563-025-00520-y.
[13] Wang J, Li Y, Wang T, Li J, Wang H, Liu P. A gold futures price forecast model based on SGRU-AM. IEEE Access. 2021;9:146745-146754. doi:10.1109/ACCESS.2021.3122140.
[14] Shahi TB, Shrestha A, Neupane A, Guo W. Stock price forecasting with deep learning: a comparative study. Mathematics. 2020;8(9):1441. doi:10.3390/math8091441.
[15] Chi DTK, Kien HNT, Nguyen TQ. Enhancing forex market forecasting with feature-augmented multivariate LSTM models using real-time data. Knowl Based Syst. 2025;330:114500. doi:10.1016/j.knosys.2025.114500.
[16] Alzakari SA, Alhussan AA, Qenawy AST, Elshewey AM, Eed M. An enhanced long short-term memory recurrent neural network deep learning model for potato price prediction. Potato Res. 2024. doi:10.1007/s11540-024-09744-x.
[17] Manogna RL, Dharmaji V, Sarang S. Enhancing agricultural commodity price forecasting with deep learning. Sci Rep. 2025;15(1):20903. doi:10.1038/s41598-025-05103-z.
[18] Xu Y, Liu T, Fang Q, Du P, Wang J. Crude oil price forecasting with multivariate selection, machine learning, and a nonlinear combination strategy. Eng Appl Artif Intell. 2025;139:109510. doi:10.1016/j.engappai.2024.109510.
[19] Zhang X, Zhang L, Zhou Q, Jin X. A novel Bitcoin and gold prices prediction method using an LSTM-P neural network model. Comput Intell Neurosci. 2022;2022:1-12. doi:10.1155/2022/1643413.
[20] Liang Y, Lin Y, Lu Q. Forecasting gold price using a novel hybrid model with ICEEMDAN and LSTM-CNN-CBAM. Expert Syst Appl. 2022;206:117847. doi:10.1016/j.eswa.2022.117847.
[21] Zhang Y, Peng Y, Song Y. Metal commodity futures price forecasting based on a hybrid secondary decomposition error-corrected model. J Big Data. 2025;12(1):166. doi:10.1186/s40537-025-01240-4.
[22] Li Y, Wang S, Wei Y, Zhu Q. A new hybrid VMD-ICSS-BiGRU approach for gold futures price forecasting and algorithmic trading. IEEE Trans Comput Soc Syst. 2021;8(6):1357-1368. doi:10.1109/TCSS.2021.3084847.
[23] Yang W, Sun S, Hao Y, Wang S. A novel machine learning-based electricity price forecasting model based on optimal model selection strategy. Energy. 2022;238:121989. doi:10.1016/j.energy.2021.121989.
[24] Dai Z, Huang H, Jiang Q, Chen Y. A parallel combined Transformer-CNN model using secondary decomposition for crude oil forecasting. Expert Syst Appl. 2026;299:129968. doi:10.1016/j.eswa.2025.129968.
[25] Zhao Y, Guo Y, Wang X. Hybrid LSTM-Transformer architecture with multi-scale feature fusion for high-accuracy gold futures price forecasting. Mathematics. 2025;13(10):1551. doi:10.3390/math13101551.
[26] Xue X, Duan P, Liu Z, Chu Q, Zhang C, Zhang B. Gated fusion enhanced multi-scale hierarchical graph convolutional network for stock movement prediction. In: Neural Information Processing (ICONIP 2025). Lecture Notes in Computer Science, vol 16311. Singapore: Springer; 2026. p. 381-395. doi:10.1007/978-981-95-4381-6_26.
[27] Arevalo J, Solorio T, Montes-y-Gómez M, González FA. Gated multimodal units for information fusion. arXiv:1702.01992 [Preprint]. 2017.
[28] Limbare A, Agarwal R. Demand forecasting and budget planning for automotive supply chain. EAI Endorsed Trans IoT. 2023;10. doi:10.4108/eetiot.4514.
Downloads
Published
Issue
Section
License
Copyright (c) 2026 Nguyen Hoang Ha, Cuong Hoa Nguyen-Dinh, Truong An Binh
This work is licensed under a Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License.
This is an open-access article distributed under the terms of the Creative Commons Attribution CC BY 4.0 license, which permits unlimited use, distribution, and reproduction in any medium so long as the original work is properly cited.
