請用此 Handle URI 來引用此文件:
http://tdr.lib.ntu.edu.tw/jspui/handle/123456789/89039
完整後設資料紀錄
DC 欄位 | 值 | 語言 |
---|---|---|
dc.contributor.advisor | 曹承礎 | zh_TW |
dc.contributor.advisor | Seng-Cho Chou | en |
dc.contributor.author | 黃柏叡 | zh_TW |
dc.contributor.author | Po-Jui Huang | en |
dc.date.accessioned | 2023-08-16T16:52:38Z | - |
dc.date.available | 2023-11-09 | - |
dc.date.copyright | 2023-08-16 | - |
dc.date.issued | 2023 | - |
dc.date.submitted | 2023-08-07 | - |
dc.identifier.citation | Amadeo Christopher, K. D. and Handoko, B. L. (2022). Forecasting cryptocurrency volatility using garch and arch model. pages 121–128.
Bansal, P. and Jain, S. (2022). Cryptocurrency price prediction using twitter and news articles analysis. pages 225–233. Boroden, C. (2008). Fibonacci Trading: How to Master the Time and Price Advantage. McGraw Hill, 1 edition. Consunji, M. P. (2021). Evotrader: Automated bitcoin trading using neuroevolutionary algorithms on technical analysis and social sentiment data. In 2021 4th International Conference on Algorithms, Computing and Artificial Intelligence, pages 1–9. Gabriel Borrageiro, N. F. and Barucca, P. (2022). The recurrent reinforcement learning crypto agent. IEEE Access, 10:38590–38599. Hado van Hasselt, A. G. and Silverg, D. (2016). Deep reinforcement learning with double q-learning. In Proceedings of the Thirtieth AAAI Conference on Artificial Intelligence (AAAI-16), pages 2094–2100. Igor Lyukevich, I. G. and Bessonova, E. (2021). Cryptocurrency market: Choice of technical indicators in trading strategies of individual investors. In Proceedings of the 3rd International Scientific Conference on Innovations in Digital Economy, pages 408–416. Jiang, Z. and Liang, J. (2017). Cryptocurrency portfolio management with deep reinforcement learnings. In Intelligent Systems Conference 2017s, pages 905–913. Khalid ABOULOULA, Ali OU-YASSINE, S.-d. K. and ELHOSENY, M. (2019). Lower the loss ratio when using a fibonacci indicator retracement to restore accuracy by avoiding wrong forecasts using enhanced indicators. In 2019 International Conference of Computer Science and Renewable Energies (ICCSRE), pages 1–8. Li, B. (2022). Markov process modeling on derived state spaces of the price dynamics of stock market indices. In Proceedings of the 2022 5th International Conference on Mathematics and Statistics, pages 66–71. Lucarelli, G. and Borrotti, M. (2019). A deep reinforcement learning approach for automated cryptocurrency trading. In Artificial Intelligence Applications and Innovations, pages 247–258. Luyao Zhang, Tianyu Wu, S. L.-C.-G. S.-F. and Li, J. (2022). A data science pipeline for algorithmic trading: A comparative study of applications for finance and cryptoeconomics. 2022 IEEE International Conference on Blockchain, pages 298–303. Micah Goldblum, Avi Schwarzschild, A. P. and Goldstein, T. (2021). Adversarial attacks on machine learning systems for high-frequency trading. pages 1–9. Panpan Wang, X. L. and Wu, S. (2022). Dynamic linkage between bitcoin and traditional financial assets: A comparative analysis of different time frequencies. volume 24, pages 1565–1586. ping Li, G. (2020). Futures trading strategy test based on computer programming. In Proceedings of the 3rd International Conference on Data Science and Information Technology, pages 159–164. Timothy P. Lillicrap, Jonathan J. Hunt, A. P. N. H. T. E. Y. T. D. S. and Wierstra, D. (2016). Continuous control with deep reinforcement learning. In ICLR 2016, volume 6, pages 1–14. Tommy Wijaya Sagala, Mei Silviana Saputri, R. M. and Budi, I. (2020). Stock price movement prediction using technical analysis and sentiment analysis. In Proceedings of the 2020 2nd Asia Pacific Information Technology Conference, pages 123–127. Tschorsch, F. and Scheuermann, B. (2016). Bitcoin and beyond: A technical survey on decentralized digital currencies. volume 18, pages 2084–2123. Yan Pang, G. S. and Ren, J. (2019). Cryptocurrency price prediction using time series and social sentiment data. pages 35–41. zhen Zhu, Z. (2022). Research on quantitative timing trading strategy based on deep reinforcement learning. pages 633–638. Çağrı Karahan, and Şule Gündüz Öğüdücü (2022). Cryptocurrency trading based on heuristic guided approach with feature engineering. pages 1–6. | - |
dc.identifier.uri | http://tdr.lib.ntu.edu.tw/jspui/handle/123456789/89039 | - |
dc.description.abstract | 近年來,隨著人工智能和機器學習技術的發展,算法交易在金融市場中變得越來越普遍。算法交易通過利用電腦程序對市場數據進行分析和預測,可以實現更快速、更準確的交易決策,並且能夠避免人為情感和主觀判斷對交易的影響。然而,傳統的算法交易方法通常基於統計模型和價格模式,無法捕捉到市場中複雜的非線性關係。技術分析作為一種常用的交易分析方法,基於市場價格和交易量等數據,通過對圖表和指標的分析,試圖預測未來價格的走勢;有鑑於此,本實驗旨在將算法交易中常用的強化學習模型結合技術分析的相關指標,期盼能在加密貨幣市場中取得更好的績效。
首先,我們將介紹實驗中會提到的一些相關背景知識,包含加密貨幣市場概況、技術分析常用指標、強化學習等概念。接下來,我們提出實驗中會用到的強化學習模型以及經過篩選後的技術分析指標,將技術分析相關指標結合到強化學習模型中。我們將使用過去的歷史市場數據來訓練強化學習模型,並使用技術分析相關指標來生成特徵向量。通過將這些特徵向量作為狀態信息,我們的強化學習模型可以學習到更有效的交易策略,並根據當前市場環境做出相應的交易決策。在實驗部分,我們將使用比特幣與以太坊的歷史數據來評估我們提出的方法。我們將比較使用技術分析相關指標的強化學習模型和常被交易新手使用的多項交易策略。我們將考慮交易的回報率、風險指標和穩定性等方面的績效指標,以評估我們的方法的優劣。最後,我們將討論實驗結果並提出未來的研究方向,並為相關領域的學術研究提供新的思路和方法。 | zh_TW |
dc.description.abstract | In recent years, with the development of artificial intelligence and machine learning technologies, algorithmic trading has become increasingly common in the financial markets. Algorithmic trading utilizes computer programs to analyze and predict market data, enabling faster and more accurate trading decisions while avoiding the influence of human emotions and subjective judgments. However, traditional algorithmic trading methods are often based on statistical models and price patterns, which fail to capture the complex nonlinear relationships in the market. Technical analysis, as a commonly used trading analysis method, is based on market price and trading volume data, attempting to predict future price trends through chart and indicator analysis. Therefore, this experiment aims to combine the commonly used reinforcement learning models in algorithmic trading with the relevant indicators of technical analysis, hoping to achieve better performance in the cryptocurrency market.
Firstly, we will introduce some background knowledge relevant to the experiment, including an overview of the cryptocurrency market, commonly used technical analysis indicators, and reinforcement learning concepts. Next, we propose the reinforcement learning model and the selected technical analysis indicators that will be used in the experiment to integrate the relevant indicators of technical analysis into the reinforcement learning model. We will train the reinforcement learning model using historical market data and generate feature vectors using the technical analysis indicators. By using these feature vectors as state information, our reinforcement learning model can learn more effective trading strategies and make corresponding trading decisions based on the current market conditions. In the experimental part, we will use historical data of Bitcoin and Ethereum to evaluate our proposed methods. We will compare the reinforcement learning model using technical analysis indicators with multiple trading strategies commonly used by novice traders. We will consider performance indicators such as return rate, risk indicators, and stability to assess the effectiveness of our methods. Finally, we will discuss the experimental results and propose future research directions, providing new ideas and methods for academic research in related fields. | en |
dc.description.provenance | Submitted by admin ntu (admin@lib.ntu.edu.tw) on 2023-08-16T16:52:38Z No. of bitstreams: 0 | en |
dc.description.provenance | Made available in DSpace on 2023-08-16T16:52:38Z (GMT). No. of bitstreams: 0 | en |
dc.description.tableofcontents | 口試委員會審定書 i
誌謝 ii 摘要 iii Abstract iv List of Figures viii List of Tables ix Chapter 1 Introduction 1 1.1 Background 1 1.2 Cryptocurrency Market Situation 2 1.3 Methods for Forecasting Price 2 1.4 Previous Algorithmic Trading Approaches for Cryptocurrency Price Forecasting 3 1.5 Purpose of this Research 4 1.6 Expected Contributions 4 Chapter 2 Literature Review 6 2.1 Technical Analysis 6 2.2 Reinforcement Learning 8 Chapter 3 Methodology 10 3.1 Problem Description 10 3.2 Double Deep Q-Network 10 3.3 Deep Deterministic Policy Gradient model 12 3.4 Input Feature 14 3.4.1 Fibonacci Price 14 3.4.2 Exponential Moving Average 15 3.4.3 BIAS 16 3.4.4 Trading Volume 16 3.4.5 Open Interest 16 3.4.6 Long/Short Ratio 17 Chapter 4 Experiments 18 4.1 Evaluation Metrics 18 4.1.1 Return of Investment 19 4.1.2 Sharpe Ratio 19 4.1.3 Sortino Ratio 19 4.2 Strategies Comparison 20 4.2.1 Relative Strength Index 20 4.2.2 Stochastic Oscillator 20 4.2.3 Moving Average Convergence Divergence 21 4.2.4 Buy and Hold 21 4.3 Experiment Details 21 4.3.1 Double Deep Q-Network 21 4.3.2 Deep Deterministic Policy Gradient 22 4.3.3 Fibonacci Selling Strategy 22 4.3.4 Benchmarks 23 4.4 Experient Result 24 4.4.1 Results on Bitcoin 2022 24 4.4.2 Results on Bitcoin 2023 Jan to May 27 4.4.3 Results on Ethereum 2022 29 4.4.4 Results on Ethereum 2023 Jan to May 31 4.5 Model Evaluation 33 4.5.1 Results on Bitcoin 2021 33 4.6 Discussions 35 4.6.1 DDPG & DDQN 35 4.6.2 Fibonacci Selling Strategy 36 Chapter 5 Conclusion and Further Work 37 References 39 | - |
dc.language.iso | en | - |
dc.title | 基於技術分析的強化學習在比特幣和以太坊上的算法交易 | zh_TW |
dc.title | Algorithmic Trading on Bitcoin and Ethereum Based on Technical Analysis by Reinforcement Learning | en |
dc.type | Thesis | - |
dc.date.schoolyear | 111-2 | - |
dc.description.degree | 碩士 | - |
dc.contributor.oralexamcommittee | 陳建錦;林俊叡 | zh_TW |
dc.contributor.oralexamcommittee | Chien-Chin Chen;Raymund Lin | en |
dc.subject.keyword | 強化學習,算法交易,技術分析,比特幣,以太坊, | zh_TW |
dc.subject.keyword | Reinforcement Learning,Algorithmic Trading,Technical Analysis,Bitcoin,Ethereum, | en |
dc.relation.page | 41 | - |
dc.identifier.doi | 10.6342/NTU202303268 | - |
dc.rights.note | 同意授權(限校園內公開) | - |
dc.date.accepted | 2023-08-09 | - |
dc.contributor.author-college | 管理學院 | - |
dc.contributor.author-dept | 資訊管理學系 | - |
顯示於系所單位: | 資訊管理學系 |
文件中的檔案:
檔案 | 大小 | 格式 | |
---|---|---|---|
ntu-111-2.pdf 授權僅限NTU校內IP使用(校園外請利用VPN校外連線服務) | 2.32 MB | Adobe PDF | 檢視/開啟 |
系統中的文件,除了特別指名其著作權條款之外,均受到著作權保護,並且保留所有的權利。