ABSTRACT
Understanding the variations in trading price (volatility), and its response to exogenous information, is a well-researched topic in finance. In this study, we focus on finding stable and accurate volatility predictors for a relatively new asset class of cryptocurrencies, in particular Bitcoin, using deep learning representations of public social media data obtained from Twitter. For our experiments, we extracted semantic information and user statistics from over 30 million Bitcoin-related tweets, in conjunction with 15-minute frequency price data over a horizon of 144 days. Using this data, we built several deep learning architectures that utilized different combinations of the gathered information. For each model, we conducted ablation studies to assess the influence of different components and feature sets over the prediction accuracy. We found statistical evidences for the hypotheses that: (i) temporal convolutional networks perform significantly better than both classical autoregressive models and other deep learning-based architectures in the literature, and (ii) tweet author meta-information, even detached from the tweet itself, is a better predictor of volatility than the semantic content and tweet volume statistics. We demonstrate how different information sets gathered from social media can be utilized in different architectures and how they affect the prediction results. As an additional contribution, we make our dataset public for future research.
Supplemental Material
- Michel Rauchs and Garrick Hileman. Global Cryptocurrency Benchmarking Study. Cambridge Centre for Alternative Finance, Cambridge Judge Business School, University of Cambridge, 2017.Google Scholar
- Triple-A. Crypto ownership table, 2020.Google Scholar
- Yuexin Mao, Wei Wei, Bing Wang, and Benyuan Liu. Correlating S&P 500 stocks with Twitter data. In Proceedings of the First ACM International Workshop on Hot Topics on Interdisciplinary Social Networks Research, HotSocial '12, page 69--72, New York, NY, USA, 2012. Association for Computing Machinery.Google ScholarDigital Library
- Svitlana Galeshchuk, O. Vasylchyshyn, and Andriy Krysovatyy. Bitcoin response to Twitter sentiments. In ICTERI Workshops, 2018.Google Scholar
- Ioannis Livieris, Niki Kiriakidou, Stavros Stavroyiannis, and P. Pintelas. An advanced CNN-LSTM model for cryptocurrency forecasting. Electronics, 01 2021.Google Scholar
- Thársis T.P. Souza and Tomaso Aste. Predicting future stock market structure by combining social and financial network information. Physica A: Statistical Mechanics and its Applications, 535:122343, Dec 2019.Google ScholarCross Ref
- Juan Piñeiro, M. López-Cabarcos, Ada M. Pérez-Pico, and Belén Ribeiro-Navarrete. Does social network sentiment influence the relationship between the S&P 500 and gold returns? International Review of Financial Analysis, 57, 02 2018.Google Scholar
- Jia-Yen Huang and Jin-Hao Liu. Using social media mining technology to improve stock price forecast accuracy. Journal of Forecasting, 39(1):104--116, 2020.Google ScholarCross Ref
- Dehua Shen, Andrew Urquhart, and Pengfei Wang. Does Twitter predict Bitcoin? Economics Letters, 174:118--122, 2019.Google ScholarCross Ref
- Dibakar Raj Pant, Prasanga Neupane, Anuj Poudel, Anup Kumar Pokhrel, and Bishnu Kumar Lama. Recurrent neural network based bitcoin price prediction by twitter sentiment analysis. In 2018 IEEE 3rd International Conference on Computing, Communication and Security (ICCCS), pages 128--132, 2018.Google ScholarCross Ref
- Suhwan Ji, Jongmin Kim, and Hyeonseung Im. A comparative study of Bitcoin price prediction using deep learning. Mathematics, 7:898, 09 2019.Google Scholar
- S. Howison, Avraam Rafailidis, and Henrik Rasmussen. On the pricing and hedging of volatility derivatives. Applied Mathematical Finance, 11:317--346, 12 2004.Google ScholarCross Ref
- Irena Barja?i? and Nino Antulov-Fantulin. Time-varying volatility in bitcoin market and information flow at minute-level frequency. Frontiers in Physics, 9:273, 2021.Google Scholar
- Alla A. Petukhina, Raphael C. G. Reule, and Wolfgang Karl Härdle. Rise of the machines? Intraday high-frequency trading patterns of cryptocurrencies. The European Journal of Finance, 27(1--2):8--30, Jul 2020.Google Scholar
- C. Hutto and Eric Gilbert. Vader: A parsimonious rule-based model for sentiment analysis of social media text. Proceedings of the International AAAI Conference on Web and Social Media, 8(1):216--225, May 2014.Google ScholarCross Ref
- HE Stanley and RN Mantegna. An introduction to econophysics. Cambridge University Press, Cambridge, 2000.Google Scholar
- Rama Cont. Empirical properties of asset returns: stylized facts and statistical issues. Quantitative finance, 1(2):223, 2001.Google ScholarCross Ref
- Robert F Engle. Autoregressive conditional heteroscedasticity with estimates of the variance of united kingdom inflation. Econometrica: Journal of the Econometric Society, pages 987--1007, 1982.Google ScholarCross Ref
- Tim Bollerslev. Generalized autoregressive conditional heteroskedasticity. Journal of econometrics, 31(3):307--327, 1986.Google ScholarCross Ref
- Philippe Jorion. On jump processes in the foreign exchange and stock markets. The Review of Financial Studies, 1(4):427--445, 1988.Google ScholarCross Ref
- Wing H Chan and John M Maheu. Conditional jump dynamics in stock market returns. Journal of Business & Economic Statistics, 20(3):377--389, 2002.Google ScholarCross Ref
- Daniele Mastro. Forecasting realized volatility: Arch-type models vs. the har-rv model. SSRN Electronic Journal, 01 2014.Google Scholar
- Yang Liu. Novel volatility forecasting using deep learning--long short term memory recurrent neural networks. Expert Systems with Applications, 132:99--109, 2019.Google ScholarDigital Library
- Nghia Nguyen, Minh-Ngoc Tran, David Gunawan, and Robert Kohn. A long short-term memory stochastic volatility model. arXiv preprint arXiv:1906.02884, 2019.Google Scholar
- Sean McNally, Jason Roche, and Simon Caton. Predicting the price of Bitcoin using machine learning. In 2018 26th euromicro international conference on parallel, distributed and network-based processing (PDP), pages 339--343. IEEE, 2018.Google Scholar
- Muhammad Saad, Jinchun Choi, DaeHun Nyang, Joongheon Kim, and Aziz Mohaisen. Toward characterizing blockchain-based cryptocurrencies for highly accurate predictions. IEEE Systems Journal, 14(1):321--332, 2019.Google ScholarCross Ref
- Qiutong Guo, Shun Lei, Qing Ye, and Zhiyang Fang. MRC-LSTM: A hybrid approach of multi-scale residual CNN and LSTM to predict Bitcoin price, 2021.Google Scholar
- Andrea Bucci. Realized volatility forecasting with neural networks. Journal of Financial Econometrics, 18(3):502--531, 2020.Google ScholarCross Ref
- Ajitha Vejendla and David Enke. Evaluation of garch, rnn and fnn models for forecasting volatility in the financial markets. IUP Journal of Financial Risk management, 10(1), 2013.Google Scholar
- Adam Hayes. Cryptocurrency value formation: An empirical analysis leading to a cost of production model for valuing Bitcoin. SSRN Electronic Journal, 2015.Google ScholarCross Ref
- Wilko Bolt. On the value of virtual currencies. SSRN Electronic Journal, 2016.Google Scholar
- Saralees Nadarajah and Jeffrey Chu. On the inefficiency of Bitcoin. Economics Letters, 150:6--9, 2017.Google ScholarCross Ref
- Eng-Tuck Cheah and John Fry. Speculative bubbles in Bitcoin markets? An empirical investigation into the fundamental value of Bitcoin. Economics Letters, 130:32--36, 2015.Google ScholarCross Ref
- Ladislav and Kristoufek. What are the main drivers of the Bitcoin price? Evidence from wavelet coherence analysis. PloS One, 10(4):e0123923, 2015.Google ScholarCross Ref
- Thomas Dimpfl and Franziska J Peter. Nothing but noise? price discovery across cryptocurrency exchanges. Journal of Financial Markets, 54:100584, 2021.Google ScholarCross Ref
- Jonathan Donier and Jean-Philippe Bouchaud. Why do markets crash? Bitcoin data offers unprecedented insights. PloS One, 10:1--11, 2015.Google ScholarCross Ref
- Paraskevi Katsiampa. Volatility estimation for Bitcoin: A comparison of GARCH models. Economics Letters, 158:3--6, 2017.Google ScholarCross Ref
- Tian Guo, Albert Bifet, and Nino Antulov-Fantulin. Bitcoin volatility forecasting with a glimpse into buy and sell orders. In 2018 IEEE International Conference on Data Mining (ICDM), pages 989--994, Nov 2018.Google ScholarCross Ref
- Nino Antulov-Fantulin, Tian Guo, and Fabrizio Lillo. Temporal mixture ensemble models for probabilistic forecasting of intraday cryptocurrency volume. Decisions in Economics and Finance, pages 1--36, 2021.Google ScholarCross Ref
- Dorit Ron and Adi Shamir. Quantitative analysis of the full Bitcoin transaction graph. In International Conference on Financial Cryptography and Data Security, pages 6--24. Springer, 2013.Google ScholarCross Ref
- Abeer ElBahrawy, Laura Alessandretti, Anne Kandler, Romualdo Pastor-Satorras, and Andrea Baronchelli. Evolutionary dynamics of the cryptocurrency market. Royal Society Open Science, 4(11):170623, 2017.Google ScholarCross Ref
- Nino Antulov-Fantulin, Dijana Tolic, Matija Piskorec, Zhang Ce, and Irena Vodenska. Inferring short-term volatility indicators from the Bitcoin blockchain. In Complex Networks and Their Applications VII, pages 508--520, Cham, 2019. Springer International Publishing.Google ScholarCross Ref
- David Garcia and Frank Schweitzer. Social signals and algorithmic trading of Bitcoin. Royal Society Open Science, 2(9):150288, 2015.Google ScholarCross Ref
- Asim K. Dey, Cuneyt G. Akcora, Yulia R. Gel, and Murat Kantarcioglu. On the role of local blockchain network features in cryptocurrency price formation. Canadian Journal of Statistics, 48(3):561--581, 2020.Google ScholarCross Ref
- Johannes Beck, Roberta Huang, David Lindner, Tian Guo, Zhang Ce, Dirk Helbing, and Nino Antulov-Fantulin. Sensing social media signals for cryptocurrency news. In Companion proceedings of the 2019 World Wide Web conference, pages 1051--1054, 2019.Google ScholarDigital Library
- Larry G. Epstein and Martin Schneider. Ambiguity, information quality, and asset pricing. The Journal of Finance, 63(1):197--228, 2008.Google ScholarCross Ref
- Qi Cao, Huawei Shen, Keting Cen, Wentao Ouyang, and Xueqi Cheng. Deephawkes: Bridging the gap between prediction and understanding of information cascades. In Proceedings of the 2017 ACM on Conference on Information and Knowledge Management, pages 1149--1158, 2017.Google ScholarDigital Library
- Hongyuan Mei and Jason M Eisner. The neural hawkes process: A neurally self-modulating multivariate point process. Advances in neural information processing systems, 30, 2017.Google Scholar
- James D. Hamilton and Princeton University Press. Time Series Analysis. Number 10. c. in Time Series Analysis. Princeton University Press, 1994.Google Scholar
- Torben G. Andersen, Tim Bollerslev, Francis X. Diebold, and Paul Labys. Modeling and forecasting realized volatility. Econometrica, 71(2):579--625, 2003.Google ScholarCross Ref
- Shaojie Bai, J Zico Kolter, and Vladlen Koltun. An empirical evaluation of generic convolutional and recurrent networks for sequence modeling. arXiv preprint arXiv:1803.01271, 2018.Google Scholar
- Junyoung Chung, Caglar Gulcehre, KyungHyun Cho, and Yoshua Bengio. Empirical evaluation of gated recurrent neural networks on sequence modelling, 2014.Google Scholar
- Colin Lea, Rene Vidal, Austin Reiter, and Gregory D Hager. Temporal convolutional networks: A unified approach to action segmentation. In European Conference on Computer Vision, pages 47--54. Springer, 2016.Google ScholarCross Ref
- Emre Aksan and Otmar Hilliges. STCN: Stochastic temporal convolutional networks. arXiv preprint arXiv:1902.06568, 2019.Google Scholar
- Aaron van den Oord, Sander Dieleman, Heiga Zen, Karen Simonyan, Oriol Vinyals, Alex Graves, Nal Kalchbrenner, Andrew Senior, and Koray Kavukcuoglu. Wavenet: A generative model for raw audio. arXiv preprint arXiv:1609.03499, 2016.Google Scholar
- Takuya Akiba, Shotaro Sano, Toshihiko Yanase, Takeru Ohta, and Masanori Koyama. Optuna: A next-generation hyperparameter optimization framework. In Proceedings of the 25rd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, 2019.Google ScholarDigital Library
- Ilya Loshchilov and Frank Hutter. Decoupled weight decay regularization. arXiv preprint arXiv:1711.05101, 2017.Google Scholar
- Martina Matta, Ilaria Lunesu, and Michele Marchesi. Bitcoin spread prediction using social and web search media. In UMAP workshops, pages 1--10, 2015.Google Scholar
- Jacob Devlin, Ming-Wei Chang, Kenton Lee, and Kristina Toutanova. BERT: Pre-training of deep bidirectional transformers for language understanding. arXiv preprint arXiv:1810.04805, 2018.Google Scholar
Index Terms
- Ask "Who", Not "What": Bitcoin Volatility Forecasting with Twitter Data
Recommendations
Bitcoin price forecasting: A perspective of underlying blockchain transactions
Highlights- We study Bitcoin price forecasting from the perspective of the underlying blockchain transactions.
- We propose a novel framework to examine the volatility of Bitcoin prices and forecast its price at different time scales.
- We find ...
AbstractCryptocurrency price forecasting plays an important role in financial markets. Traditional approaches face two challenges: (1) it is difficult to ascertain the influential factors related to price forecasting; and (2) due to the 24/7 trading ...
Volatility Spillovers between US Banking Industry and Bitcoin Market: Risk Implications for Banking Industry
ICBCT'20: Proceedings of the 2020 The 2nd International Conference on Blockchain TechnologyThis paper examines the volatility spillovers between Bitcoin market and US banking industry using unrestricted BEKK-GARCH model. The results show that there is a strong short-term volatility spillover effect in the two markets. However, Bitcoin trading ...
Information resonance on Twitter: watching Iran
SOMA '10: Proceedings of the First Workshop on Social Media AnalyticsTwitter has undoubtedly caught the attention of both the general public, and academia as a microblogging service worthy of study and attention. Twitter has several features that sets it apart from other social media/networking sites, including its 140 ...
Comments