پژوهش ها و چشم اندازهای اقتصادی

پژوهش ها و چشم اندازهای اقتصادی

تحلیل روند و پیش‌بینی قیمت بیت کوین ‌ یک تحلیل مقایسه ای از ‌مدل‌های خطی و غیر ‌خطی ‌

مقالات آماده انتشار
نویسندگان
مینو نظیفی نایینی 1
رسول بخشی دستجردی 2
سید عباس حسینی غفار 3
1 دانشجوی دکتری اقتصاد مالی ، گروه علوم اقتصادی، دانشکدۀ اقتصاد، دانشگاه شهید اشرفی اصفهانی، اصفهان، ایران
2 دانشیار، گروه علوم اقتصادی، دانشکدۀ اقتصاد، دانشگاه اصفهان، اصفهان، ایران
3 استادیار ، گروه علوم اقتصادی، دانشکدۀ اقتصاد، دانشگاه شهید اشرفی اصفهانی، اصفهان، ایران
10.48311/ecor.2025.24053
چکیده
تحلیل و پیش‌بینی قیمت بیت‌کوین به دلیل ماهیت غیرخطی، نوسان‌پذیر و تأثیرپذیر از عوامل گوناگون، همواره چالشی مهم در حوزه مالی و اقتصادی به شمار می‌رود. این پژوهش با هدف بررسی روند قیمتی و پیش‌بینی نوسانات بیت‌کوین، از داده‌های روزانه بین سال‌های ۲۰۱۴ تا ۲۰۲۵ بهره گرفته و مجموعه‌ای از متغیرهای فنی (حجم معاملات، سختی شبکه، میزان ذخیره بیت‌کوین)، اقتصادی (شاخص S&P500، نرخ بهره آمریکا، قیمت نفت، شاخص دلار) و اجتماعی (تعداد سرچ گوگل و توییت های بیت‌کوین)را وارد مدل کرده است.

در بخش روش‌شناسی، از سه رویکرد مکمل استفاده شده است:


شبکه عصبی مصنوعی (ANN) جهت پیش‌بینی دقیق قیمت‌ها،


۲. رگرسیون چرخشی مارکف (MSR) برای شناسایی رژیم‌های نوسانی بازار و اثر شوک‌ها،

۳. ابزارهای تحلیل تکنیکال، ایچیموکو(RSI) برای تفسیر روندهای بازار.

یافته‌ها حاکی از آن است که شبکه عصبی پیش‌بینی دقیقی ارائه می‌دهد و در آن، شاخص S&P بیشترین اثر را دارد. مدل مارکف نشان می‌دهد که بازار بیت‌کوین در بلندمدت بیشتر در وضعیت کم‌نوسان قرار دارد، و نرخ بهره آمریکا و EUR/USD از متغیرهای کلیدی‌اند. همچنین، تحلیل تکنیکال روند کلی صعودی را تأیید می‌کند، اما RSI و ابر ایچیموکو هشدار اصلاحات کوتاه‌مدت را می‌دهند.

نتایج پژوهش، کاربرد راهبردی برای تصمیم‌گیری سرمایه‌گذاران بلندمدت و معامله‌گران کوتاه‌مدت دارد و ضرورت توجه به متغیرهای کلان اقتصادی در کنار شاخص‌های فنی را برجسته می‌سازد.

کلیدواژه‌ها
بیتکوین
شبکه عصبی
پیش‌بینی
‌اندیکاتور
ایچیموکو
بازار‌‌‌های مالی

موضوعات

عنوان مقاله English

Analyzing and Forecasting Bitcoin Price Trends: A Comparative Study of Linear and Nonlinear Models

نویسندگان English

mino nazifi naeini 1
rasool bakhshi dastjerdi 2
E. ABBAS HSSEINI GHAFFAR 3
1 Ph.D. Student in Financial Economics, Department of Economic Sciences, Faculty of Economics, Shahid Ashrafi Isfahani University, Isfahan, Iran
2 Associate Professor, Department of Economic Sciences, Faculty of Economics, Isfahan University, Isfahan, Iran
3 Assistant Professor, Department of Economic Sciences, Faculty of Economics, Shahid Ashrafi Isfahani University, Isfahan, Iran
چکیده English

Financial markets, including the Bitcoin market, are inherently complex and nonlinear. Predicting their direction and strength remains a significant challenge for traders and investors. This study aims to model the strength of Bitcoin price trends using a Markov convolutional regression model, artificial neural networks (ANN), and time series analysis. To evaluate trend strength, technical indicators and oscillators such as the Relative Strength Index (RSI) and the Ichimoku trading system are employed. Daily Bitcoin price data from January 1, 2014, to July 9, 2023, are analyzed. The findings indicate that the neural network model effectively forecasts Bitcoin prices, while the Markov convolutional regression model successfully identifies periods of rotation and shock. Furthermore, the analysis reveals that the Bitcoin market experiences longer durations of low volatility compared to high volatility periods. Among the influencing factors in the ANN model, the Standard & Poor’s 500 Index has the most substantial impact on Bitcoin prices, surpassing other economic variables. Notably, the time series model demonstrates the lowest mean absolute error. Across all three models, network difficulty, global gold prices, and exchange rates significantly affect Bitcoin prices.

Aim and Introduction

Amid declining trust in traditional fiat currencies, Bitcoin has emerged as a potential safe haven and a store of value. Its adoption may increase over time, potentially positioning it as a global currency or an asset for wealth preservation. However, like other financial markets, the Bitcoin market is characterized by inherent complexity and nonlinearity, making price trend analysis and prediction particularly challenging for investors and traders. These complexities necessitate the development of advanced predictive models to capture the dynamic behavior of Bitcoin prices.

Methodology

The primary objective of this study is to analyze and forecast Bitcoin price trends using three distinct modeling approaches: the Markov convolutional regression model, artificial neural networks, and time series analysis. To assess the strength of price trends, technical indicators and oscillators such as the RSI, MACD, and the Ichimoku trading system are utilized.

The study employs both technical and economic variables. Technical variables include the number of daily Bitcoin transactions, network difficulty, Bitcoin reserves, and transaction volumes. Economic variables include the S&P 500 Index, the New York Stock Exchange Index, the U.S. Dollar Index, U.S. interest rates, and oil prices. Data were collected on a daily basis from July 9, 2014, to July 9, 2023. The research design is applied in nature, and the statistical population comprises Bitcoin price data. The econometric approach incorporates the comparative evaluation of the three models, particularly focusing on their accuracy in trend prediction and their responsiveness to various economic and technical indicators.

Results and Discussion

The results demonstrate that the artificial neural network model outperforms the other two models in predicting Bitcoin prices. In particular, the S&P 500 Index emerges as the most influential variable in the ANN model, highlighting a strong linkage between traditional equity markets and cryptocurrency behavior.

Both the RSI and Ichimoku indicators suggest a probable decrease in Bitcoin prices in the upcoming months, followed by a potential increase toward the end of 2024. The Markov convolutional regression model effectively identifies years marked by market rotation and shocks and reveals that periods of low volatility tend to last longer than those of high volatility.

Furthermore, the time series model, while not outperforming the ANN in accuracy, exhibits the lowest mean absolute error. Across all models, the variables of network difficulty, global gold price, and exchange rate consistently show statistically significant impacts on Bitcoin pricing, underscoring their crucial role in shaping market dynamics.

Conclusion

The comparative analysis of the three modeling approaches reveals that the artificial neural network model yields the most accurate forecasts of Bitcoin prices, as indicated by its lowest mean squared error. The consistent significance of the network difficulty, global gold price, and exchange rate variables across models reflects growing investor confidence in Bitcoin as a financial asset. This suggests the potential for capital movement from traditional markets, such as gold and Forex, toward cryptocurrencies.

Additionally, the application of MACD and Ichimoku systems confirms the anticipated short-term decline and subsequent long-term increase in Bitcoin prices. These findings offer valuable insights for policymakers, investors, and analysts aiming to understand the dynamics of cryptocurrency markets and improve predictive accuracy

کلیدواژه‌ها English

Bitcoin
Financial markets
forecasting
Ichimoku
indicator
Neural Network
Time Series
Abraham, R (2020). The role of investor sentiment in the valuation of ‎bitcoin and bitcoin derivatives. International Journal of Financial Markets ‎and Derivatives. https://doi.org/10.1504/ijfmd.2020.10031383‎
Aggarwal, A et al. (2019). Deep Learning Approach to Determine the ‎Impact of Socio Economic Factors on Bitcoin Price Prediction. 2019 ‎Twelfth International Conference on Contemporary Computing (IC3), 1-5. ‎https://doi.org/10.1109/IC3.2019.8844928‎
Aghashahi, M, Bamdad, S (2022). Analysis of different artificial neural ‎networks for Bitcoin price prediction. International Journal of Management ‎Science and Engineering Management, 18, 126 - 133. ‎https://doi.org/10.1080/17509653.2022.2032442‎
Almansour, B et al. (2020). The Impact Of Exchange Rates On Bitcoin ‎Returns: Further Evidence From A Time Series Framework. ‎https://www.semanticscholar.org/paper/ef83b403c0687ff36a7f2665c766b11‎‎7eef64699‎
Aoyagi, J, Hattori, T (2019). The Empirical Analysis of Bitcoin Market in ‎the General Equilibrium Framework. Banking & Insurance eJournal. ‎https://doi.org/10.2139/ssrn.3433833‎
Aragon, R G M, Linares, A A A (2023). Political-economic uncertainty and ‎the price of Bitcoin in the forex Market Between 2010 and 2020. Ciencia ‎Latina Revista Científica Multidisciplinar. ‎https://doi.org/10.37811/cl_rcm.v7i2.5626‎
Athey, S et al. (2016). Bitcoin Pricing, Adoption, and Usage: Theory and ‎Evidence. Stanford Graduate School of Business Research Paper Series. ‎https://www.semanticscholar.org/paper/8a7c799b5734dfd961606ace218deab‎192e06ec5‎
Ayaz, Z et al. (2020). Bitcoin Price Prediction using ARIMA Model. ‎https://doi.org/10.36227/techrxiv.12098067‎
Azari, A (2019). Bitcoin Price Prediction: An ARIMA Approach. ArXiv, ‎abs/1904.05315. ‎https://www.semanticscholar.org/paper/4c5bdca2bda17627f00847ea9f79b77‎c750a24e2‎
Banjade, D (2020). Forecasting Bitcoin Price using Artificial Neural ‎Network. Capital Markets: Asset Pricing & Valuation eJournal. ‎https://doi.org/10.2139/ssrn.3515702‎
Barro, R. J. (1979). Money and the price level under the gold standard. The ‎Economic Journal, 89(353), 13 -33. DOR: ‎‎20.1001.1.26453355.1399.13.46.1.6‎
Bartoš, J (2015). Does Bitcoin follow the hypothesis of efficient market. ‎https://doi.org/10.20472/ES.2015.4.2.002‎
Chen, J (2023). Analysis of Bitcoin Price Prediction Using Machine ‎Learning. Journal of Risk and Financial Management. ‎https://doi.org/10.3390/jrfm16010051‎
Chennupati, A et al. (2024). Comparative Analysis of Bitcoin Price ‎Prediction Models: LSTM, BiLSTM, ARIMA and Transformers. 2024 15th ‎International Conference on Computing Communication and Networking ‎Technologies (ICCCNT), 1-7. ‎https://doi.org/10.1109/ICCCNT61001.2024.10724247‎
Ciaian, P et al. (2014). The economics of BitCoin price formation. Applied ‎Economics, 48, 1799 - 1815. ‎https://doi.org/10.1080/00036846.2015.1109038‎
Dai, M et al. (2022). An Assessment of the Impact of Natural Resource ‎Price and Global Economic Policy Uncertainty on Financial Asset ‎Performance: Evidence From Bitcoin. ‎https://doi.org/10.3389/fenvs.2022.897496‎
Deng, W (2023). Analysis of the Factors Affecting the Price Fluctuation of ‎Bitcoin. Advances in Economics, Management and Political Sciences. ‎https://doi.org/10.54254/2754-1169/3/2022832‎
Drozdz S, Minati L, Oswiecimka P, Stanuszek M, Watorek M (2019) ‎Signatures of the crypto-currency market decoupling from the forex. ‎Future Internet 11(7):154. https://doi.org/10.3390/fi11070154‎
Dubey, P (2022). Short-run and long-run determinants of bitcoin returns: ‎transnational evidence. Review of Behavioral Finance. ‎https://doi.org/10.1108/rbf-02-2022-0040‎
Dutta, A A et al. (2019). A Gated Recurrent Unit Approach to Bitcoin ‎Price Prediction. Economics of Innovation eJournal. ‎https://doi.org/10.3390/jrfm13020023‎
Enders, W. (2015). Applied econometric time series fourth edition.‎
Erfanian, S et al. (2022). Predicting Bitcoin (BTC) Price in the Context of ‎Economic Theories: A Machine Learning Approach. Entropy, 24. ‎https://doi.org/10.3390/e24101487‎
Fantazzini, D, Kolodin, N (2020). Does the Hashrate Affect the Bitcoin ‎Price? Information Systems eJournal. https://doi.org/10.2139/ssrn.3720039‎
García, D, Schweitzer, F (2015). Social signals and algorithmic trading of ‎Bitcoin. Royal Society Open Science, 2. https://doi.org/10.1098/rsos.150288‎
Gozbasi, O et al. (2021). IS BITCOIN A SAFE HAVEN? A STUDY ON ‎THE FACTORS THAT AFFECT BITCOIN PRICES. International Journal ‎of Economics and Financial Issues. https://doi.org/10.32479/IJEFI.11602‎
Gronwald, M (2014). The Economics of Bitcoins -- Market Characteristics ‎and Price Jumps. ERN: Asset Pricing Models (Topic) ‎https://doi.org/10.2139/ssrn.2548999‎
Hartono, D J, Suyanto, S (2023). Major determinants of Bitcoin price: ‎Application of a vector error correction model. Investment Management ‎and Financial Innovations. https://doi.org/10.21511/imfi.20(4).2023.21‎
Hassan, M M (2024). Bitcoin Price Prediction Using Deep Bayesian LSTM ‎With Uncertainty Quantification: A Monte Carlo Dropout–Based ‎Approach. Stat. https://doi.org/10.1002/sta4.70001‎
Hoang, L T, Baur, D (2021). Effects of Bitcoin Exchange Reserves on ‎Bitcoin Returns and Volatility. Investments eJournal. ‎https://doi.org/10.2139/ssrn.3902504‎
Huang, W et al. (2022). Time Series Analysis and Prediction on Bitcoin. ‎BCP Business & Management. https://doi.org/10.54691/bcpbm.v34i.3163‎
Ji, S et al. (2019). A Comparative Study of Bitcoin Price Prediction Using ‎Deep Learning. Mathematics. https://doi.org/10.3390/math7100898‎
Kamboj, D et al. (2023). A Comparative Study of Time Series Models for ‎Bitcoin Price Prediction. 2023 14th International Conference on ‎Computing Communication and Networking Technologies (ICCCNT), 1-5. ‎https://doi.org/10.1109/ICCCNT56998.2023.10306819‎
Kasimati, A E et al. (n.d.). AIS Electronic. ‎https://www.semanticscholar.org/paper/98cf2e0427fa75cb7a1ba55a9c909b4‎‎3d6f6a151‎
Khadija, M et al. (2024). Prediction of Bitcoin Prices Based on Blockchain ‎Information: A Deep Reinforcement Learning Approach. Adv. Artif. Intell. ‎Mach. Learn., 4, 2416-2433. https://doi.org/10.54364/aaiml.2024.43141‎
Khedmati, M et al. (2020). Time Series Forecasting of Bitcoin Price Based ‎on Autoregressive Integrated Moving Average and Machine Learning ‎Approaches. International Journal of Engineering. ‎https://doi.org/10.5829/ije.2020.33.07a.16‎
Kjærland, F et al. (2018). An Analysis of Bitcoin’s Price Dynamics. Journal ‎of Risk and Financial Management. https://doi.org/10.3390/JRFM11040063‎
Koroleva, E et al. (2024). The impact of socio-economic and technical ‎factors on the price of Bitcoin. Financial Analytics: Science and ‎Experience. https://doi.org/10.24891/fa.17.2.162‎
Kristoufek, L. (2013). BitCoin meets Google Trends and Wikipedia: ‎Quantifying the relationship between phenomena of the Internet era. ‎Scientific reports , 3, 3415.‎
Lamothe-Fernández, P et al. (2020). Deep Learning Methods for Modeling ‎Bitcoin Price. Mathematics. https://doi.org/10.3390/math8081245‎
Li, X, Wang, C (2014). The Technology and Economic Determinants of ‎Cryptocurrency Exchange Rates: The Case of Bitcoin. IRPN: Innovation & ‎Finance (Topic) https://doi.org/10.2139/ssrn.2515233‎
Lu, T (2023). US Monetary Policy Uncertainty and Changes in ‎Cryptocurrency Market: Evidence from ARIMA Model. Highlights in ‎Business, Economics and Management. ‎https://doi.org/10.54097/hbem.v20i.12331‎
Marthinsen, J, Gordon, S (2022). The Price and Cost of Bitcoin. The ‎Quarterly Review of Economics and Finance. ‎https://doi.org/10.1016/j.qref.2022.04.003‎
Metaxas, D N (2024). ANALYSIS OF THE EFFECT OF EXCHANGE ‎RATE FLUCTUATIONS ON BITCOIN RETURNS: A TIME SERIES ‎APPROACH IN A DIGITAL FINANCE PERSPECTIVE. Join: Journal of ‎Social Science. https://doi.org/10.59613/fgc4tj101‎
Mohammadjafari, A (2024). Comparative Study of Bitcoin Price Prediction. ‎ArXiv, abs/2405.08089. https://doi.org/10.48550/arXiv.2405.08089‎
Mungoli N. (2023.) Deciphering the Blockchain: A Comprehensive ‎Analysis of Bitcoin's Evolution. Adoption, and Future Implications.‎
Nasution, M R et al. (2023). THE EFFECT OF TECHNICAL ASSET, ‎FINANCIAL ASSET AND MACRO-ECONOMIC ON RETURN BITCOIN. ‎International Journal of Economic, Business, Accounting, Agriculture ‎Management and Sharia Administration (IJEBAS) ‎https://doi.org/10.54443/ijebas.v3i2.728‎
Nepp, A, Karpeko, F (2022). Hype as a Factor on the Global Market: The ‎Case of Bitcoin. Journal of Behavioral Finance, 25, 1 - 14. ‎https://doi.org/10.1080/15427560.2022.2073593‎
Noda, S et al. (2019). An Economic Analysis of Difficulty Adjustment ‎Algorithms in Proof-of-Work Blockchain Systems. Proceedings of the 21st ‎ACM Conference on Economics and Computation. ‎https://doi.org/10.2139/ssrn.3410460‎
Orăștean, R et al. (2024). Exploring the Relationship Between Google ‎Trends and Cryptocurrency Metrics. Studies in Business and Economics, ‎‎19, 368 - 379. https://doi.org/10.2478/sbe-2024-0020‎
Prates, R, Fonseca, M W d (2024). Do News Reports and Google Searches ‎Impact Bitcoin Prices? An Analysis of Granger Causality. Revista de ‎Gestão Social e Ambiental. https://doi.org/10.24857/rgsa.v18n10-165‎
Samadi . s, Nazifi Naeini.M ,(2013) "Analysis the Factors Affecting Gold ‎Price Fluctuations Using a Markov Switching Regression Models and ‎Neural Networks",Volume 20, Issue 6 - Serial Number 6 August ‎‎2013,Pages 121-146, https://doi.org/10.22067/pm.v20i6.36359(in Persian) ‎
Serafini, G et al. (2020). Sentiment-Driven Price Prediction of the Bitcoin ‎based on Statistical and Deep Learning Approaches. 2020 International ‎Joint Conference on Neural Networks (IJCNN), 1-8. ‎https://doi.org/10.1109/IJCNN48605.2020.9206704‎
Su, C et al. (2022). CAN BITCOIN BE A SAFE HAVEN IN FEAR ‎SENTIMENT? Technological and Economic Development of Economy. ‎https://doi.org/10.3846/tede.2022.15502‎
Suhwan JiJongmin KimHyeonseung Im,2019,A Comparative Study of ‎Bitcoin Price Prediction Using Deep Learning
Umeokwobi, R et al. (2024). Nexus between Monetary Indicators and ‎Bitcoin in Selected Sub-Saharan Africa: A Panel ARDL. Theoretical and ‎Practical Research in Economic Fields. ‎https://doi.org/10.14505/tpref.v15.3(31).18‎
Velankar, S. S. Valecha and S. Maji (2018). Bitcoin price prediction using ‎machine learning, 20th International Conference on Advanced ‎Communication Technology (ICACT).‎
Wahyuni, M T et al. (2024). US macroeconomic determinants of Bitcoin. ‎Investment Management and Financial Innovations. ‎https://doi.org/10.21511/imfi.21(2).2024.19‎
Wang, J et al. (2016). An Analysis of Bitcoin Price Based on VEC Model. ‎https://doi.org/10.2991/ICEMI-16.2016.36‎
Wang, X et al. (2020). The Relationship Between Bitcoin and Stock ‎Market. Int. J. Oper. Res. Inf. Syst., 11, 22-35. ‎https://doi.org/10.4018/ijoris.2020040102‎
Wu, R et al. (2024). Factors affecting the volatility of bitcoin prices. The ‎Economics and Finance Letters. https://doi.org/10.18488/29.v11i2.3730‎
Zhao, C (2024). Research on prediction of bitcoin price based on machine ‎learning methods. SHS Web of Conferences. ‎https://doi.org/10.1051/shsconf/202418102006‎
Zhu, M (2023). Research on the Impact Mechanism of Economic Policy ‎Uncertainty on Bitcoin Prices. Frontiers in Business, Economics and ‎Management. https://doi.org/10.54097/fbem.v10i1.10236‎
Zhu, Y et al. (2023). Price Prediction of Bitcoin Based on Adaptive Feature ‎Selection and Model Optimization. Mathematics. ‎https://doi.org/10.3390/math11061335‎
Zuo, A (2024). Comparison and Analysis of the Accuracy of Various ‎Machine Learning Algorithms in Bitcoin Price Prediction. Advances in ‎Economics, Management and Political Sciences. ‎https://doi.org/10.54254/2754-1169/70/20231726‎
‎ ‎
پژوهش ها و چشم اندازهای اقتصادی

مقالات آماده انتشار، پذیرفته شده
انتشار آنلاین از 25 شهریور 1404