Bitcoin Price Volatility Prediction on Technical Indicators with GARCH and LSTM
Abstract
Bitcoin is one of the most volatile cryptocurrency assets, infl uenced by technical factors, market sentiment, and macroeconomic conditions. This high volatility poses both challenges in risk management and opportunities for market participants. This study aims to predict Bitcoin price volatility by developing a hybrid GARCH-LSTM model that combines the strengths of statistical approaches and deep learning. Historical Bitcoin data from June 2009 to June 2025 was collected through scraping from TradingView and enriched with seven technical indicators (SMA, EMA, RSI, Stochastic Oscillator, OBV, and MFI). Dimensionality reduction using PCA produced two principal components explaining 82.9% of the variance. Stationarity was confi rmed using the ADF test, while GARCH (1,2) was selected based on AIC and BIC criteria to capture short-term volatility patterns. GARCH outputs then integrated into LSTM to learn long-term non-linear patterns. Model performance was evaluated using MAE, RMSE, and MAPE. Results indicate that the hybrid GARCH (1,1)-LSTM model achieved the best performance, with MAE = 0.24981, RMSE = 0.39597, and MAPE = 7.8%, demonstrating high accuracy for highly volatile cryptocurrency data. Although long-term forecast accuracy declined, this model shows strong potential for applications in Value-at-Risk strategies, short-term trading decisions, and asset portfolio allocation. Future research is recommended to incorporate external variables such as market sentiment and macroeconomic factors to enhance adaptability to dynamic market conditions.
References
Nakamoto S. Bitcoin: A Peer-to-Peer Electronic Cash System [Internet]. 2008. Available from: www.bitcoin.org
Peng S, Prentice C, Shams S, Sarker T. A systematic literature review on the determinants of cryptocurrency pricing. China Accounting and Finance Review. 2024 Mar 5;26(1):1–30.
Sapkota N. News-based sentiment and bitcoin volatility. International Review of Financial Analysis. 2022 Jul 1;82.
Huang JZ, Huang W, Ni J. Predicting bitcoin returns using high-dimensional technical indicators. Journal of Finance and Data Science. 2019 Sep 1;5(3):140–55.
Fazlollahi N, Ebrahimijam S. Predicting Cryptocurrency Price Returns by Using Deep Learning Model of Technical Analysis Indicators. In: Springer Proceedings in Business and Economics. Springer Nature; 2023. p. 175–86.
Napolès G, Nicenboim B. Predicting Bitcoin price using technical indicator data features in Long short-term Memory models. 2021.
Lubis JK, Kharisudin I. Metode Long Short Term Memory dan Generalized Autoregressive Conditional Heteroscedasticity untuk Pemodelan Data Saham. PRISMA, Prosiding Seminar Nasional Matematika [Internet]. 2021;4:652–8. Available from: https://journal.unnes.ac.id/sju/index.php/prisma/
Robert Kissell. Price Volatility. In: The Science of Algorithmic Trading and Portfolio Management. Elsevier Inc.; 2013. p. 193–234.
Poon SH, Granger CWJ. Forecasting Volatility in Financial Markets: A Review. Vol. XLI, Journal of Economic Literature. 2003.
Liem JH, Calvin D, Tjong C, Juan K:, Liem H. Penerapan CRISP-DM untuk Prediksi Harga Saham NVIDIA Menggunakan Time Series Analysis. 2024;4:204–13.
Salma Maataoui, Ghita Bencheikh, Ghizlane Bencheikh. 2023 Fifth International Conference on Advances in Computational Tools for Engineering Applications (ACTEA). IEEE; 2023. 44 p.
I T Jolliffe. Principal Component Analysis, Second Edition. 2002.
Bollerslev T. GENERALIZED AUTOREGRESSIVE CONDITIONAL HETEROSKEDASTICITY. Vol. 31, Journal of Econometrics. 1986.
Mualifah LNA, Soleh AM, Notodiputro KA. Comparison of GARCH, LSTM, and Hybrid GARCH-LSTM Models for Analyzing Data Volatility. International Journal of Advances in Soft Computing and its Applications. 2024;16(2):150–65.
Li JH, Guo SX, Ma RL, He J, Zhang XH, Rui DS, et al. Comparison of the effects of imputation methods for missing data in predictive modelling of cohort study datasets. BMC Med Res Methodol. 2024 Dec 1;24(1).
J. Welles Wilder Jr. New Concepts in Technical Trading Systems. 1978.
Gerald Appel. Technical Analysis - Power Tools For Active Investors [Internet]. 1979. Available from: www.pearsoned-ema.com.
Willa Dhany H, Izhari F. Journal of Intelligent Decision Support System (IDSS) Exploratory Data Analysis (EDA) methods for healthcare classification. Vol. 6, Journal of Intelligent Decision Support System (IDSS). 2023.
Agarwal A, Prabha S, Yadav R. Exploratory Data Analysis (EDA) for Banking and Finance: Unveiling Insights and Patterns. 2024.
Sung SH, Kim JM, Park BK, Kim S. A Study on Cryptocurrency Log-Return Price Prediction Using Multivariate Time-Series Model. Axioms. 2022 Sep 1;11(9).
Nur IM, Nugrahanto R, Fauzi F. CRYPTOCURRENCY PRICE PREDICTION: A HYBRID LONG SHORT-TERM MEMORY MODEL WITH GENERALIZED AUTOREGRESSIVE CONDITIONAL HETEROSCEDASTICITY. BAREKENG: Jurnal Ilmu Matematika dan Terapan. 2023 Sep 30;17(3):1575–84.
Shen Z, Wan Q, Leatham DJ. Bitcoin Return Volatility Forecasting: A Comparative Study between GARCH and RNN. Journal of Risk and Financial Management. 2021 Jul 1;14(7).
Pujitha MV, Kumar KS, Sree DI, Devi PY. Predicting Bitcoin Prices for the Next 30 Days Using LSTM-Based Time Series Analysis. In: 2023 2nd International Conference on Futuristic Technologies, INCOFT 2023. Institute of Electrical and Electronics Engineers Inc.; 2023.
Wang Y, Liu P, Zhu K, Liu L, Zhang Y, Xu G. A Garlic-Price-Prediction Approach Based on Combined LSTM and GARCH-Family Model. Applied Sciences (Switzerland). 2022 Nov 1;12(22).
