學術產出-學位論文
文章檢視/開啟
書目匯出
-
題名 運用LSTM進行Bitcoin價格預測
Applying LSTM to Bitcoin price prediction作者 陳維睿
Chen, Wei-Rui貢獻者 胡毓忠
Hu, Yuh-Jong
陳維睿
Chen, Wei-Rui關鍵詞 長短期記憶
比特幣
區塊鏈
Long Short-Term Memory
Bitcoin
Blockchain日期 2018 上傳時間 2-八月-2018 16:22:44 (UTC+8) 摘要 本論文運用長短期記憶模型(Long Short-Term Memory, LSTM) 來預測比特幣(Bitcoin)價格走向。特徵值資料包含內部及外部特徵值，各抽取自比特幣區塊鏈以及交易中心。\n\n加密貨幣是一種新型態的貨幣，其交易運行在網路中。在所有加密貨幣中，比特幣(Bitcoin, BTC)是第一個加密貨幣，且目前擁有最高的市值。預測比特幣價格是一個新興的研究題目，因為其與傳統金融資產有所差異，且其價格非常波動。\n\n本論文對比特幣區塊鏈資料處理方法提出指引，並將長短期記憶模型實務應用到比特幣價格預測。
This thesis focuses on applying Long Short-Term Memory (LSTM) technique to predict Bitcoin price direction. Features including internal and external features are extracted from Bitcoin blockchain and exchange center respectively.\n\nCryptocurrency is a new type of currency that is traded over the infrastructure of Internet. Bitcoin (BTC) is the first cryptocurrency and ranks first in the market capitalization among all the other cryptocurrencies. Predicting Bitcoin price is a novel topic because of its differences with traditional financial assets and its volatility.\n\nAs contributions, this thesis provides a guide of processing Bitcoin blockchain data and serves as an empirical study on applying LSTM to Bitcoin price prediction.參考文獻 1] S. Hochreiter and J. Schmidhuber, “Long short-term memory,” Neural Computation,\nvol. 9, no. 8, 1997.\n[2] S. Nakamoto, “Bitcoin: A peer-to-peer electronic cash system,” 2009.\n[3] I. Goodfellow, Y. Bengio, and A. Courville, Deep Learning.\nMIT Press, 2016.\n[4] Y. LeCun, Y. Bengio, and G. Hinton, “Deep learning,” Nature, vol. 521, 2015.\n[5] F. A. Gers, J. A. Schmidhuber, and F. A. Cummins, “Learning to forget: Continual\nprediction with lstm,” Neural Comput., vol. 12, no. 10, 2000.\n[6] S. J. Taylor, An Introduction to Volatility. Princeton University Press, 2005.\n[7] Investopedia. Volatility. [Online]. Available: https://www.investopedia.com/terms/\nv/volatility.asp\n[8] W. Huang, Y. Nakamori, and S.-Y. Wang, “Forecasting stock market movement\ndirection with support vector machine,” Computers & Operations Research, vol. 32,\nno. 10, 2005.\n[9] S. A. Hamid and Z. Iqbal, “Using neural networks for forecasting volatility of sp 500\nindex futures prices,” Journal of Business Research, 2004.\n[10] A. Vejendla and D. Enke, “Evaluation of garch, rnn and fnn models for forecasting\nvolatility in the financial markets,” IUP Journal of Financial Risk Management,\nvol. 10, no. 1, 2013.\n[11] R. Akita, A. Yoshihara, T. Matsubara, and K. Uehara, “Deep learning for stock\nprediction using numerical and textual information,” in 2016 IEEE/ ACIS 15th\nInternational Conference on Computer and Information Science (ICIS), 2016.\n[12] M. Matta, M. I. Lunesu, and M. Marchesi, “Bitcoin spread prediction using social\nand web search media,” in UMAP Workshops, 2015.\n[13] I. Madan and S. Saluja, “Automated bitcoin trading via machine learning\nalgorithms,” Stanford University, 2014.\n[14] A. Greaves and B. Au, “Using the bitcoin transaction graph to predict the price of\nbitcoin,” Stanford University, 2015.\n[15] S. McNally, “Predicting the price of bitcoin using machine learning,” Master’s thesis,\nDublin, National College of Ireland, 2016.\n[16] H. Jang and J. Lee, “An empirical study on modeling and prediction of bitcoin prices\nwith bayesian neural networks based on blockchain information,” IEEE Access, vol. 6,\n2018.\n[17] Y. Bengio, “Learning deep architectures for ai,” Foundations and Trends® in Machine\nLearning, vol. 2, no. 1, 2009.\n[18] Y. Bengio, A. Courville, and P. Vincent, “Representation learning: A review and\nnew perspectives,” IEEE Transactions on Pattern Analysis and Machine Intelligence,\nvol. 35, no. 8, 2013.\n[19] J. L. Elman, “Finding structure in time,” Cognitive Science, vol. 14, no. 2, 1990.\n[20] Z. C. Lipton, “A critical review of recurrent neural networks for sequence learning,”\nCoRR, vol. abs/1506.00019, 2015.\n[21] A. Graves, Supervised Sequence Labelling with Recurrent Neural Networks.\nSpringer-Verlag Berlin Heidelberg, 2012.\n[22] K. Greff, R. K. Srivastava, J. Koutník, B. R. Steunebrink, and J. Schmidhuber,\n“LSTM: A search space odyssey,” CoRR, vol. abs/1503.04069, 2015.\n[23] Wikipedia contributors, “Loss functions for classification — Wikipedia, the free\nencyclopedia,” 2018. [Online]. Available:\nhttps://en.wikipedia.org/w/index.php?\ntitle=Loss_functions_for_classification&oldid=838253245\n[24] Wikipedia contributors, “Gradient descent — Wikipedia, the free encyclopedia,”\n2018. [Online]. Available: https://en.wikipedia.org/w/index.php?title=Gradient_\ndescent&oldid=845809247\n[25] R. Rojas, Neural Networks: A Systematic Introduction.\nBerlin, Heidelberg:\nSpringer-Verlag, 1996.\n[26] J. Bergstra and Y. Bengio, “Random search for hyper-parameter optimization,” J.\nMach. Learn. Res., vol. 13, 2012.\n[27] N. Srivastava, G. Hinton, A. Krizhevsky, I. Sutskever, and R. Salakhutdinov,\n“Dropout: A simple way to prevent neural networks from overfitting,” Journal of\nMachine Learning Research, vol. 15, 2014.\n[28] S. Ruder, “An overview of gradient descent optimization algorithms,” CoRR, vol.\nabs/1609.04747, 2016.\n[29] D. P. Kingma and J. Ba, “Adam: A method for stochastic optimization,” CoRR, vol.\nabs/1412.6980, 2014.\n[30] S. Dziembowski, “Introduction to cryptocurrencies,” 2015.\n[31] I. Bentov, A. Gabizon, and A. Mizrahi, “Cryptocurrencies without proof of work,”\nCoRR, vol. abs/1406.5694, 2014.\n[32] Proof of work. [Online]. Available: https://en.bitcoin.it/wiki/Proof_of_work\n[33] A. Narayanan, J. Bonneau, E. W. Felten, A. Miller, S. Goldfeder, and J. Clark,\nBitcoin and Cryptocurrency Technologies. Princeton University Press, 2016.\n[34] Gdax exchange center documentation. [Online]. Available: https://docs.gdax.com/\n[35] blockchain.info. [Online]. Available: https://blockchain.info/\n[36] T. Hastie, R. Tibshirani, and J. Friedman, The Elements of Statistical Learning.\nSpringer New York Inc., 2001.\n[37] Keras. [Online]. Available: https://keras.io/\n[38] Nvidia. [Online]. Available: http://www.nvidia.com/page/home.html\n[39] A. Karpathy, “The unreasonable effectiveness of recurrent neural networks,” 2015.\n[Online]. Available: http://karpathy.github.io/2015/05/21/rnn-effectiveness/ 描述 碩士
國立政治大學
資訊科學系
105753015資料來源 http://thesis.lib.nccu.edu.tw/record/#G0105753015 資料類型 thesis dc.contributor.advisor 胡毓忠 zh_TW dc.contributor.advisor Hu, Yuh-Jong en_US dc.contributor.author (作者) 陳維睿 zh_TW dc.contributor.author (作者) Chen, Wei-Rui en_US dc.creator (作者) 陳維睿 zh_TW dc.creator (作者) Chen, Wei-Rui en_US dc.date (日期) 2018 en_US dc.date.accessioned 2-八月-2018 16:22:44 (UTC+8) - dc.date.available 2-八月-2018 16:22:44 (UTC+8) - dc.date.issued (上傳時間) 2-八月-2018 16:22:44 (UTC+8) - dc.identifier (其他 識別碼) G0105753015 en_US dc.identifier.uri (URI) https://ah.lib.nccu.edu.tw/item?item_id=137073 - dc.description (描述) 碩士 zh_TW dc.description (描述) 國立政治大學 zh_TW dc.description (描述) 資訊科學系 zh_TW dc.description (描述) 105753015 zh_TW dc.description.abstract (摘要) 本論文運用長短期記憶模型(Long Short-Term Memory, LSTM) 來預測比特幣(Bitcoin)價格走向。特徵值資料包含內部及外部特徵值，各抽取自比特幣區塊鏈以及交易中心。\n\n加密貨幣是一種新型態的貨幣，其交易運行在網路中。在所有加密貨幣中，比特幣(Bitcoin, BTC)是第一個加密貨幣，且目前擁有最高的市值。預測比特幣價格是一個新興的研究題目，因為其與傳統金融資產有所差異，且其價格非常波動。\n\n本論文對比特幣區塊鏈資料處理方法提出指引，並將長短期記憶模型實務應用到比特幣價格預測。 zh_TW dc.description.abstract (摘要) This thesis focuses on applying Long Short-Term Memory (LSTM) technique to predict Bitcoin price direction. Features including internal and external features are extracted from Bitcoin blockchain and exchange center respectively.\n\nCryptocurrency is a new type of currency that is traded over the infrastructure of Internet. Bitcoin (BTC) is the first cryptocurrency and ranks first in the market capitalization among all the other cryptocurrencies. Predicting Bitcoin price is a novel topic because of its differences with traditional financial assets and its volatility.\n\nAs contributions, this thesis provides a guide of processing Bitcoin blockchain data and serves as an empirical study on applying LSTM to Bitcoin price prediction. en_US dc.description.tableofcontents 1 Introduction 1\n1.1 Research Objective 1\n1.2 Deep Learning on Time Series data 1\n1.3 Predicting Bitcoin Price 2\n1.4 Related Works 4\n\n2 LSTM on Time Series Data 5\n2.1 Neural Network and Deep Learning 5\n2.2 Recurrent Neural Network 6\n2.3 Long Short-Term Memory 7\n2.4 Training an LSTM Network 9\n2.4.1 Loss Function 9\n2.4.2 Gradient Descent 10\n2.4.3 Backpropagation and Backpropagation Through Time 11\n2.4.4 Hyperparamter tuning 11\n2.4.4.1 Dropout Rate 12\n2.4.4.2 Neural Network Optimization Algorithm 12\n\n3 Bitcoin and Blockchain 14\n3.1 Bitcoin on Blockchain 14\n3.2 Bitcoin as a Cryptocurrency 15\n3.3 Mining Bitcoin 17\n\n4 Machine Learning Pipeline\n4.1 Pipeline 19\n4.2 Data Collection 20\n4.2.1 Collecting Bitcoin Blockchain Data 20\n4.2.2 Collecting Data from Exchange Center 21\n4.3 Data Cleaning 21\n4.4 Data Processing 22\n4.4.1 Internal Features Extraction 22\n4.4.2 External Features Extraction 22\n4.4.3 Align and Combine Internal and External Features 23\n4.4.4 Min-Max Normalization 23\n4.4.5 Train/Validation/Test Split 24\n\n5 Methodology 25\n5.1 Tools and Platform 25\n5.2 Experiments 25\n5.2.1 Dataset Summary 25\n5.2.2 Neural Network Architecture 26\n5.2.3 Hyperparameter Tuning 28\n5.3 Results 30\n5.3.1 Performance Comparison with Related Works 30\n\n6 Conclusion and Future Work 32\n6.1 Conclusion 32\n6.2 Future work 33\n\nAppendices 34\nA 34\nB 36\n\nReferences 39 zh_TW dc.format.extent 755725 bytes - dc.format.mimetype application/pdf - dc.source.uri (資料來源) http://thesis.lib.nccu.edu.tw/record/#G0105753015 en_US dc.subject (關鍵詞) 長短期記憶 zh_TW dc.subject (關鍵詞) 比特幣 zh_TW dc.subject (關鍵詞) 區塊鏈 zh_TW dc.subject (關鍵詞) Long Short-Term Memory en_US dc.subject (關鍵詞) Bitcoin en_US dc.subject (關鍵詞) Blockchain en_US dc.title (題名) 運用LSTM進行Bitcoin價格預測 zh_TW dc.title (題名) Applying LSTM to Bitcoin price prediction en_US dc.type (資料類型) thesis en_US dc.relation.reference (參考文獻) 1] S. Hochreiter and J. Schmidhuber, “Long short-term memory,” Neural Computation,\nvol. 9, no. 8, 1997.\n[2] S. Nakamoto, “Bitcoin: A peer-to-peer electronic cash system,” 2009.\n[3] I. Goodfellow, Y. Bengio, and A. Courville, Deep Learning.\nMIT Press, 2016.\n[4] Y. LeCun, Y. Bengio, and G. Hinton, “Deep learning,” Nature, vol. 521, 2015.\n[5] F. A. Gers, J. A. Schmidhuber, and F. A. Cummins, “Learning to forget: Continual\nprediction with lstm,” Neural Comput., vol. 12, no. 10, 2000.\n[6] S. J. Taylor, An Introduction to Volatility. Princeton University Press, 2005.\n[7] Investopedia. Volatility. [Online]. Available: https://www.investopedia.com/terms/\nv/volatility.asp\n[8] W. Huang, Y. Nakamori, and S.-Y. Wang, “Forecasting stock market movement\ndirection with support vector machine,” Computers & Operations Research, vol. 32,\nno. 10, 2005.\n[9] S. A. Hamid and Z. Iqbal, “Using neural networks for forecasting volatility of sp 500\nindex futures prices,” Journal of Business Research, 2004.\n[10] A. Vejendla and D. Enke, “Evaluation of garch, rnn and fnn models for forecasting\nvolatility in the financial markets,” IUP Journal of Financial Risk Management,\nvol. 10, no. 1, 2013.\n[11] R. Akita, A. Yoshihara, T. Matsubara, and K. Uehara, “Deep learning for stock\nprediction using numerical and textual information,” in 2016 IEEE/ ACIS 15th\nInternational Conference on Computer and Information Science (ICIS), 2016.\n[12] M. Matta, M. I. Lunesu, and M. Marchesi, “Bitcoin spread prediction using social\nand web search media,” in UMAP Workshops, 2015.\n[13] I. Madan and S. Saluja, “Automated bitcoin trading via machine learning\nalgorithms,” Stanford University, 2014.\n[14] A. Greaves and B. Au, “Using the bitcoin transaction graph to predict the price of\nbitcoin,” Stanford University, 2015.\n[15] S. McNally, “Predicting the price of bitcoin using machine learning,” Master’s thesis,\nDublin, National College of Ireland, 2016.\n[16] H. Jang and J. Lee, “An empirical study on modeling and prediction of bitcoin prices\nwith bayesian neural networks based on blockchain information,” IEEE Access, vol. 6,\n2018.\n[17] Y. Bengio, “Learning deep architectures for ai,” Foundations and Trends® in Machine\nLearning, vol. 2, no. 1, 2009.\n[18] Y. Bengio, A. Courville, and P. Vincent, “Representation learning: A review and\nnew perspectives,” IEEE Transactions on Pattern Analysis and Machine Intelligence,\nvol. 35, no. 8, 2013.\n[19] J. L. Elman, “Finding structure in time,” Cognitive Science, vol. 14, no. 2, 1990.\n[20] Z. C. Lipton, “A critical review of recurrent neural networks for sequence learning,”\nCoRR, vol. abs/1506.00019, 2015.\n[21] A. Graves, Supervised Sequence Labelling with Recurrent Neural Networks.\nSpringer-Verlag Berlin Heidelberg, 2012.\n[22] K. Greff, R. K. Srivastava, J. Koutník, B. R. Steunebrink, and J. Schmidhuber,\n“LSTM: A search space odyssey,” CoRR, vol. abs/1503.04069, 2015.\n[23] Wikipedia contributors, “Loss functions for classification — Wikipedia, the free\nencyclopedia,” 2018. [Online]. Available:\nhttps://en.wikipedia.org/w/index.php?\ntitle=Loss_functions_for_classification&oldid=838253245\n[24] Wikipedia contributors, “Gradient descent — Wikipedia, the free encyclopedia,”\n2018. [Online]. Available: https://en.wikipedia.org/w/index.php?title=Gradient_\ndescent&oldid=845809247\n[25] R. Rojas, Neural Networks: A Systematic Introduction.\nBerlin, Heidelberg:\nSpringer-Verlag, 1996.\n[26] J. Bergstra and Y. Bengio, “Random search for hyper-parameter optimization,” J.\nMach. Learn. Res., vol. 13, 2012.\n[27] N. Srivastava, G. Hinton, A. Krizhevsky, I. Sutskever, and R. Salakhutdinov,\n“Dropout: A simple way to prevent neural networks from overfitting,” Journal of\nMachine Learning Research, vol. 15, 2014.\n[28] S. Ruder, “An overview of gradient descent optimization algorithms,” CoRR, vol.\nabs/1609.04747, 2016.\n[29] D. P. Kingma and J. Ba, “Adam: A method for stochastic optimization,” CoRR, vol.\nabs/1412.6980, 2014.\n[30] S. Dziembowski, “Introduction to cryptocurrencies,” 2015.\n[31] I. Bentov, A. Gabizon, and A. Mizrahi, “Cryptocurrencies without proof of work,”\nCoRR, vol. abs/1406.5694, 2014.\n[32] Proof of work. [Online]. Available: https://en.bitcoin.it/wiki/Proof_of_work\n[33] A. Narayanan, J. Bonneau, E. W. Felten, A. Miller, S. Goldfeder, and J. Clark,\nBitcoin and Cryptocurrency Technologies. Princeton University Press, 2016.\n[34] Gdax exchange center documentation. [Online]. Available: https://docs.gdax.com/\n[35] blockchain.info. [Online]. Available: https://blockchain.info/\n[36] T. Hastie, R. Tibshirani, and J. Friedman, The Elements of Statistical Learning.\nSpringer New York Inc., 2001.\n[37] Keras. [Online]. Available: https://keras.io/\n[38] Nvidia. [Online]. Available: http://www.nvidia.com/page/home.html\n[39] A. Karpathy, “The unreasonable effectiveness of recurrent neural networks,” 2015.\n[Online]. Available: http://karpathy.github.io/2015/05/21/rnn-effectiveness/ zh_TW dc.identifier.doi (DOI) 10.6814/THE.NCCU.CS.005.2018.B02 -
