Comparative Analysis of Temporal Fusion Transformer and Long Short-Term Memory Architecture Resilience in Predicting Solana Price Volatility Across Different Market Phases

Mahdy Eka Putra; Tanty  Oktavia

doi:10.52436/1.jutif.2026.7.3.5894

Authors

Mahdy Eka Putra Master of Information Systems Management, Bina Nusantara University, Jakarta, Indonesia
Tanty Oktavia Master of Information Systems Management, Bina Nusantara University, Jakarta, Indonesia

DOI:

https://doi.org/10.52436/1.jutif.2026.7.3.5894

Keywords:

Cryptocurrency, Deep Learning, Long Short-Term Memory, Solana, Temporal Fusion Transformer, Volatility

Abstract

Abstract must be written in English. The high volatility of cryptocurrency markets, particularly for altcoins like Solana (SOL), presents a significant challenge for predictive modeling. Traditional deep learning architectures often struggle to adapt to sudden market regime shifts. Therefore, this study aims to provide a comparative analysis of the resilience between the Temporal Fusion Transformer and Long Short-Term Memory architectures in predicting Solana price volatility across three distinct market phases: the bull market of 2024, the bear market of 2025, and the recovery phase of 2026. We utilized hourly historical price and volume data combined with technical indicators such as Relative Strength Index (RSI). The models were evaluated using Root Mean Square Error (RMSE), Mean Absolute Error (MAE), and a specific performance degradation rate formula. The results demonstrate that while LSTM performs adequately during stable trends, its accuracy degrades massively by 1575.69% during high-volatility regime changes due to memory inertia causing a severe lagging effect. Conversely, the TFT model exhibited superior resilience, limiting its performance degradation to only 218.53% during the extreme bear market phase. The inherent attention mechanism and skip connections in TFT allow it to dynamically adapt to sudden structural breaks in real-time without delay. Furthermore, the implementation of the TFT architecture proved to be 62% more computationally efficient than LSTM. This research significantly contributes to the field of computer science and informatics, specifically in adaptive time-series forecasting, by proving that attention mechanisms and skip connections can efficiently solve the memory inertia problem in recurrent networks during real-time structural breaks.

Downloads

Download data is not yet available.

References

I. Yousaf and L. Yarovaya, "Herding behavior in conventional cryptocurrency market, non-fungible tokens, and DeFi assets," Financ. Res. Lett., vol. 50, p. 103299, 2022. doi: 10.1016/j.frl.2022.103299.

M. Hidayatullah and A. Juniar, "Narrative Research Study: Market Sentiment As A Trigger For Cryptocurrency Volatility," Syntax Literate, vol. 14, no. 1, pp. 261–292, 2024. doi: 10.33506/sl.v14i1.3900.

M. L. Pratama and H. Utama, "Pendekatan Deep Learning Menggunakan Metode LSTM untuk Prediksi Harga Bitcoin," Indones. J. Comput. Sci. Res., vol. 2, no. 2, pp. 43–50, 2023. doi: 10.37905/ijcsr.v2i2.4350.

S. M. Qureshi, A. Saeed, F. Ahmad, A. R. Khattak, and S. H. Almotiri, "Evaluating machine learning models for predictive accuracy in cryptocurrency price forecasting," PeerJ Comput. Sci., vol. 11, pp. 1–54, 2025. doi: 10.7717/peerj-cs.2626.

R. Hidayat and R. D. Irawan, "Prediksi Harga Saham Syariah Menggunakan Metode Deep Learning GRU dan LSTM," J. Tekno Kompak, vol. 20, no. 1, pp. 37–50, 2022. doi: 10.33365/jtk.v20i1.474.

N. W. Saputra, F. Insani, S. Agustian, and S. Sanjaya, "Penerapan Deep Learning Menggunakan Gated Recurrent Unit Untuk Memprediksi Harga Minyak Mentah Dunia," Build. Informatics, Technol. Sci., vol. 5, no. 1, pp. 86–94, 2023. doi: 10.47065/bits.v5i1.3552.

F. N. Wahyu and S. Anggai, "Prediksi Harga Cryptocurrency Menggunakan Algoritma Temporal Fusion Transformer, N-Beats dan Deepar," Ranah Res. J., vol. 8, no. 1, pp. 967–981, 2025. doi: 10.38035/rrj.v8i1.1949.

A. Rosyd, A. I. Purnamasari, and I. Ali, "Penerapan Metode Long Short Term Memory (LSTM) dalam Memprediksi Harga Saham PT Bank Central Asia," JATI (J. Mhs. Tek. Inform.), vol. 8, no. 1, pp. 501–506, 2024. doi: 10.36040/jati.v8i1.8440.

R. Irawan and E. Utami, "Prediksi Harga Bitcoin Menggunakan Model Hibrida Transformer dan Long Short Term Memory," J. Pendidik. dan Teknol. Indones., vol. 5, no. 9, pp. 2866–2877, 2025. doi: 10.52436/1.jpti.1091.

E. Patriya and A. Latif, "Peramalan Harga Saham Penutupan Indeks Harga Saham Gabungan (IHSG) Menggunakan Algoritma Long Short Term Memory (LSTM)," J. Ilm. Ekon. Bisnis, vol. 28, no. 2, pp. 304–314, 2023. doi: 10.35760/eb.2023.v28i2.7964.

B. Lim, S. Ö. Arık, N. Loeff, and T. Pfister, "Temporal Fusion Transformers for interpretable multi-horizon time series forecasting," Int. J. Forecast., vol. 37, no. 4, pp. 1748–1764, 2021. doi: 10.1016/j.ijforecast.2021.03.012.

M. A. Akmal and A. R. Pratama, "Implementation of Temporal Fusion Transformer (TFT) for Short-Term Sales Prediction of Telkomsel Data Packages in East Java," J. TIKAR, vol. 12, no. 1, pp. 399–413, 2026. doi: 10.37012/jtik.v12i1.3268.

L. D. A. N. Gru, "Prediksi Harga Emas Menggunakan Metode LSTM dan GRU," JITET (J. Inform. dan Tek. Elektro Ter.), vol. 11, no. 3, pp. 620–627, 2023. doi: 10.23960/jitet.v11i3.3250.

I. F. Amri, S. A. Astuti, and I. Sulistiya, "Peramalan Harga Emas Antam Menggunakan Metode Generalized Autoregressive Conditional Heteroskedasticity (GARCH)," Unnes J. Math. Comput., vol. 10, no. 1, pp. 26–35, 2022. doi: 10.52166/ujmc.v10i1.4679.

I. Nurhaida, M. Sobiri, and S. Jaya, "Optimasi Prediksi Cryptocurrency Menggunakan Pendekatan Deep Learning," J. Sci. Appl. Informatics, vol. 6, no. 2, pp. 197–204, 2023. doi: 10.36085/jsai.v6i2.5288.

G. Budiprasetyo, M. Hani, and D. Zahira, "Prediksi Harga Saham Syariah Menggunakan Algoritma Long Short-Term Memory (LSTM)," J. Nas. Sist. Inf., vol. 3, no. 3, pp. 164–172, 2023. doi: 10.25077/TEKNOSI.v8i3.2022.164-172.

N. Insights et al., "Novel Insights on the Comparative Study Between LSTM and Transformer Models for Financial Time Series Prediction," J. Eng. Technol. Ind. Appl., vol. 11, no. 55, pp. 212–221, 2025. doi: 10.5935/jetia.v11i55.2658.

M. F. Rizkilloh and S. Widiyanesti, “Prediksi Harga Cryptocurrency Menggunakan Algoritma Long Short Term Memory (LSTM),” Rekayasa Sist. dan Teknol. Inf., vol. 5, no. 1, pp. 1–2, 2026. doi: 10.29207/resti.v6i1.3630.

K. Sofi, A. S. Sunge, S. R. Riady, and A. Z. Kamalia, "Perbandingan Algoritma Linear Regression, LSTM, dan GRU dalam Memprediksi Harga Saham dengan Model Time Series," in Seminastika, vol. 3, no. 1, pp. 12-20, 2021. doi: 10.47002/seminastika.v3i1.275.

A. Sujjada, F. Sembiring, and Febriansyah, "Prediksi Harga Bitcoin Menggunakan Algoritma Long Short-Term Memory," Inov. Sist. Inf., vol. 9, no. 1, pp. 450–459, 2024. doi: 10.35314/isi.v9i1.4247.

J. Yang, P. Li, Y. Cui, X. Han, and M. Zhou, "Multi-Sensor Temporal Fusion Transformer for Stock Performance Prediction: An Adaptive Sharpe Ratio Approach," Sensors, vol. 25, no. 3, pp. 1–24, 2025. doi: 10.3390/s25030976.

E. Dinçer and Z. H. Kilimci, "From LSTM to GPT-2: Recurrent and Transformer-Based Deep Learning Architectures for Multivariate High-Liquidity Cryptocurrency Price Forecasting," Symmetry, vol. 18, no. 1, pp. 1–23, 2026. doi: 10.3390/sym18010032.

J. Yang, P. Li, Y. Cui, X. Han, and M. Zhou, “Multi-Sensor Temporal Fusion Transformer for Stock Performance Prediction: An Adaptive Sharpe Ratio Approach,” Sensors, vol. 25, no. 3, pp. 1–24, 2025. doi: 10.3390/s25030976.

E. Dinçer and Z. H. Kilimci, “From LSTM to GPT-2: Recurrent and Transformer-Based Deep Learning Architectures for Multivariate High-Liquidity Cryptocurrency Price Forecasting,” Symmetry, vol. 18, no. 1, pp. 1–23, 2026. doi: 10.3390/sym18010032.

A. J. Armando, J. G. Siento, T. A. Eikwine, Diana, and I. H. Parmonangan, “Temporal Fusion Transformer for Multi Horizon Bitcoin Price Forecasting,” in Proc. Inf. Technol. Int. Semin. (ITIS), 2023. doi: 10.1109/ITIS59651.2023.10420330.

M. C. Lee, “Temporal Fusion Transformer-Based Trading Strategy for Multi-Crypto Assets Using On-Chain and Technical Indicators,” Systems, vol. 13, no. 6, p. 473, 2025. doi: 10.3390/systems13060474.