隨著區塊鏈的快速發展，加密貨幣被越來越多人關注，比特幣亦成為近年來最熱門的虛擬貨幣，許多投資者更將比特幣視為如股票般的投資工具。然而，比特幣的價格起伏劇烈，導致投資者無法精準預測比特幣的價格，進而影響投資策略。鑒於技術指標可以有效地預測股價以及特徵擷取方法有助於股價預測和破產預測，因此本研究使用24個預測股市的技術指標值作為預測比特幣的特徵，例如: 指數平滑異同移動平均線(MACD)、布林通道(Bollinger Bands)、隨機指標(KD)、威廉指標(W%R)等。接著，使用6種特徵擷取方法來挑選代表性的特徵，例如: Information gain、主成分分析(Principal components analysis)等。最後，使用8種分類演算法及5種預測演算法來預測比特幣價格趨勢及比特幣價格。分類演算法包含K-近鄰演算法(K-nearest neighbor)、支援向量機(Support vector machine)等。預測演算法包含K-近鄰演算法(K-nearest neighbor)、支援向量迴歸(Support vector regression)等。本研究亦使用移動視窗法來進行實驗，維持訓練資料及測試資料的時間相關性。本研究期望透過特徵擷取方法及機器學習方法，探究何種技術指標值、分類演算法、預測演算法及特徵擷取方法能最有效預測比特幣價格趨勢。

With the rapid development of blockchain, cryptocurrency has attracted more and more attention. Bitcoin has become the most popular virtual currency in recent years. Many investors regard bitcoin as a stock-like investment tool. However, the price of Bitcoin fluctuates drastically, the investors can not accurately predict the price of Bitcoin. In view of the fact that technical indicators can effectively predict stock prices and feature extraction methods are helpful in forecasting stock price and bankruptcy, we propose using 24 popular stock market technical indicators, e.g., Moving Average Convergence Divergence(MACD), Bollinger Bands, Stochastic Oscillator(KD), Williams %R(W%R), to predict the price of bitcoin. Next, six feature selection methods e.g., Information gain, Principal components analysis, are used to select more useful indicators. Finally, eight classification algorithms and five prediction algorithms are employed to predict the trend of bitcoin price and the bitcoin price. The experiments uses the sliding window scheme to maintain the temporal correlation of training data and test data. The experiment results show that the technical indicator can effectively predict Bitcoin price and also indicate the technical indicators and algorithms which are effective in prediction.

摘要 I
Abstract III
目錄 V
圖目錄 IX
表目錄 XI
公式目錄 XIII
第一章 緒論 1
1.1 研究背景 1
1.2 研究動機 2
1.3 研究目的 3
第二章 文獻探討 5
2.1 使用技術指標預測股價的相關文獻 5
2.2 預測比特幣的相關文獻 5
2.3 使用特徵擷取方法建立預測模型的相關文獻 7
2.4 分類模型 8
2.4.1 自適應增強(AdaBoost) 8
2.4.2 貝式網路(Bayesian Network) 9
2.4.3 決策樹(Decision Tree) 11
2.4.4 K-近鄰演算法(K-Nearest Neighbor, KNN) 12
2.4.5 邏輯迴歸(Logistic Regression) 13
2.4.6 單純貝式分類器(Naïve Bayes Classifier) 14
2.4.7 支援向量機(Support Vector Machine) 14
2.4.8 類神經網路(Neural Network, NN) 15
2.5 預測模型 16
2.5.1 線性迴歸(Linear Regression) 16
2.5.2 決策樹(Decision Tree) 17
2.5.3 K-近鄰演算法(K-Nearest Neighbor, KNN) 17
2.5.4 支援向量迴歸(Support Vector Regression, SVR) 17
2.5.5 類神經網路(Neural Network, NN) 18
第三章 研究方法 19
3.1 研究架構 19
3.2 技術指標 21
3.3 特徵擷取 31
3.3.1 皮爾森相關係數(Pearson correlation coefficient, PCC) 31
3.3.2 Correlation-based Feature Selector (CFS) 32
3.3.3 Information gain 32
3.3.4 Information gain ratio 33
3.3.5 主成分分析(Principal components analysis, PCA) 33
3.4 移動視窗法 34
第四章 實驗結果 35
4.1 實驗環境 35
4.2 實驗資料 35
4.3 資料處理 35
4.4 實驗結果 36
4.4.1 移動視窗長度 36
4.4.2 最佳參數 37
4.4.3 比特幣 38
4.4.4 比特幣特徵擷取 40
4.4.5 以太幣、萊特幣、瑞波幣 46
第五章 結論 51
參考文獻 53
附錄 57
附錄A 主成分分析相關權重-價格趨勢預測 57
附錄B 主成分分析相關權重-價格預測 59

[1] S. Nakamoto, Bitcoin: A Peer-to-Peer Electronic Cash System, 2008, https://s3.amazonaws.com/academia.edu.documents/32413652/BitCoin_P2P_electronic_cash_system.pdf?AWSAccessKeyId=AKIAIWOWYYGZ2Y53UL3A&Expires=1547189022&Signature=djU9Pp4vr%2FGHUYga3HcpeCBunlI%3D&response-content-disposition=inline%3B%20filename%3DBitcoin_A_Peer-to-Peer_Electronic_Cash_S.pdf, retrieved on 2019/1/9.
[2] R. Peachavanish, Stock Selection and Trading Based on Cluster Analysis of Trend and Momentum Indicators, Proceedings of the International MultiConference of Engineers and Computer Scientists 2016, Vol.1, 2016.
[3] Y. Zhai, A. Hsu, and S.K. Halgamuge, Combining News and Technical Indicators in Daily Stock Price Trends Prediction, 4th International Symposium on Neural Networks-ISNN 2007, pp.1087-1096, 2007.
[4] K.J. Kim, Financial time series forecasting using support vector machines, Neurocomputing, Vol.55, pp.307-319, 2003.
[5] M.C. Lee, Using support vector machine with a hybrid feature selection method to the stock trend prediction, Expert Systems with Applications, Vol.36, pp.10896-10904, 2009.
[6] R. Choudhry, and K. Garg, A Hybrid Machine Learning System for Stock Market Forecasting, Journal of International Technology and Information Management, Vol.20, 2011.
[7] A. Kar, Stock Prediction using Artificial Neural Networks, http s://pdfs.semanticscholar.org/4e59/7dfd9907e328c355a1c965fa6df51b07a932.pdf, retrieved on 2019/1/10.
[8] M. Tanaka-Yamawaki, and S. Tokuoka, Adaptive use of technical indicators for the prediction of intra-day stock prices, Physica A: Statistical Mechanics and its Applications, Vol.383, pp.125-133, 2007.
[9] Y.B. Kim, J. Lee, N. Park, J. Choo, J.H. Kim, and C.H. Kim, When Bitcoin encounters information in an online forum: Using text mining to analyse user opinions and predict value fluctuation, Research article, 2017.
[10] I. Madan, S. Saluja, and A. Zhao, Automated Bitcoin Trading via Machine Learning Algorithms, National College of Ireland, 2015.
[11] S. McNally, J. Roche, and S. Caton, Predicting the Price of Bitcoin Using Machine Learning, 2018 26th Euromicro International Conference on Parallel, Distributed and Network-based Processing, pp.339-343, 2018.
[12] H. Jang, and J. Lee, An Empirical Study on Modeling and Prediction of Bitcoin Prices With Bayesian Neural Networks Based on Blockchain Information, IEEE Access, Vol.6, pp.5427-5437, 2018.
[13] C.F. Tsai, and Y.C. Hsiao, Combining multiple feature selection methods for stock prediction: Union, intersection, and multi-intersection approaches, Decision Support Systems, Vol.50, pp.258-269, 2010.
[14] C.F. Tsai, Feature selection in bankruptcy prediction, Knowledge-Based Systems, Vol.22, pp.120-127, 2009.
[15] D. Liang, C.F. Tsai, and H.T. Wu, The effect of feature selection on financial distress prediction, Knowledge-Based Systems, Vol.73, pp.289-297, 2015.
[16] 蕭仲君, The prediction of Bitcoin price based on technical indicators, 碩士論文, 2018.
[17] 加密錢包數量統計, https://zh.wikipedia.org/wiki/%E6%AF%94%E7%89 %B9%E5%B8%81, retrieved on 2019/1/9.
[18] 有向無環圖示意圖, https://zh.wikipedia.org/wiki/%E6%9C%89%E5%90% 91%E6%97%A0%E7%8E%AF%E5%9B%BE, retrieved on 2019/1/10.
[19] 貝式網路示意圖, https://zh.wikipedia.org/wiki/File:%E5%9C%96%E4% BA%8C%EF%BC%9A%E4%B8%80%E5%80%8B%E7%B0%A1%E5%96%AE%E7%9A%84%E8%B2%9D%E6%B0%8F%E7%B6%B2%E8%B7%AF%E4%BE%8B%E5%AD%90.jpg, retrieved on 2019/1/10.
[20] 決策樹示意圖, https://www.tutorialspoint.com/data_mining/dm_dti.htm, retrieved on 2019/1/10.
[21] K-近鄰演算法示意圖, https://medium.com/@adi.bronshtein/a-quick-introduction-to-k-nearest-neighbors-algorithm-62214cea29c7, retrieved on 2019/1/10.
[22] 邏輯函數(S型函數)示意圖, https://en.wikipedia.org/wiki/File:Logistic-curve.svg, retrieved on 2019/1/10.
[23] 支援向量機示意圖, https://docs.opencv.org/2.4/doc/tutorials/ml/i ntroduction_to_svm/introduction_to_svm.html, retrieved on 2019/1/10.
[24] 類神經網路示意圖, https://medium.com/marketingdatascience/%E5%BF%AB %E9%80%9F%E5%8F%8D%E6%87%89%E6%A9%9F%E5%88%B6-%E9%A1% 9E%E7%A5%9E%E7%B6%93%E7%B6%B2%E8%B7%AF-a3bbdee4a6f6, retrieved on 2019/6/19.
[25] 線性迴歸示意圖, https://zh.wikipedia.org/wiki/%E7%B7%9A%E6%80%A7 %E5%9B%9E%E6%AD%B8#/media/File:Linear_regression.svg, retrieved on 2019/6/19.
[26] CoinGecko, https://www.coingecko.com/zh-tw, retrieved on 2019/6/19.