加密貨幣作為投資工具的可行性，是近年來相當受到關注的議題。比特幣為一種加密貨幣，不受地區的限制通用於全世界。比特幣被視為一種投資工具時，其操作方式類似於投資一般常見的外匯一樣，如：美元、歐元、英鎊，但由於比特幣的波動較傳統法定貨幣還大，因此預測比特幣的價格較傳統的法定貨幣具有跳戰性。為了預測比特幣的報酬率，本研究使用前四大發展時間較長的加密貨幣－萊特幣、以太幣、瑞波幣和達世幣－的歷史資料，以及使用文字探勘對比特幣相關線上論壇的文章，找出對於報酬率有影響的關鍵字向量，作為解釋變數。我們以比特幣日報酬率為反應變數，使用時間序列隱馬可夫模型(Hidden Markov model)預測比特幣的日報酬率走勢。對於比特幣隔天價格走勢漲跌以及幅度的預測，有助於對未來投資做決策。

The feasibility of cryptocurrency as an investment tool has been a topic of considerable concern in recent years. Bitcoin is a cryptocurrency that is universally used regardless of regional restrictions. When Bitcoin is considered as an investment tool, it operates in the same way as investing in legal currencies, such as: US Dollars, Euros, and British Pounds. However, due to the fluctuations of Bitcoin return are larger than the legal currencies, Bitcoin return forecasting is a big challenge compared to those for the legal currencies. In order to predict the return of Bitcoin, this thesis tries to build up a prediction model based on the historical data of the Bitcoin and other top five cryptocurrencies such as the Litecoin, Ethereum, Ripple, and Dash which have a longer development time. Moreover, we include the information obtained from text mining analyzing the Bitcoin online forum articles to find the keyword sentiment scores as explanatory variables into the prediction model as well. The hidden Markov model (HMM) has been used to predict the trend of the Bitcoin daily rate of return. The prediction of the ups and downs trend of Bitcoin on the next day with acceptable accuracy should be helpful for making decisions on the investment strategy of Bitcoin.

論文審定書 i
誌謝 ii
摘要 iii
Abstract iv
前言及研究動機 1
加密貨幣介紹及文獻回顧 1
歷史與發展 1
加密貨幣特色 2
比特幣 2
以太幣 3
文獻回顧 3
虛擬貨幣資料介紹 4
文字探勘 6
爬蟲 6
文本詞彙矩陣 6
TF-IDF 演算法 7
情緒詞彙表 9
通用情緒詞彙表 9
比特幣報酬率詞彙表 10
研究方法及模型 11
自迴歸差分整合移動平均-廣義自迴歸條件異變異數模型 12
自迴歸差分整合移動平均模型 12
廣義自迴歸條件異變異數模型 12
隱馬可夫模型 13
選模準則 15
赤池信息準則 15
貝式信息準則 16
方均跟差 16
研究結果 17
情緒分數 18
ARIMAX-GARCH 模型分析結果 18
隱馬可夫模型分析結果 27
制定投資策略下之投資報酬率 34
討論與結語 36
參考文獻 37

Akaike, H. (1974). ``A new look at the statistical model identification". Automatic Control, IEEE Transactions on, 19, 716–723.

Bakar, N. A., Rosbi, S. (2017). ``Autoregressive Integrated Moving Average (ARIMA) Model for Forecasting Cryptocurrency Exchange Rate in High Volatility Environment: A New Insight of Bitcoin Transaction." International Journal of Advanced Engineering Research and Science, 4, 130-137.

Bollerslev, T (1986).``Generalized Autoregressive Conditional Heteroskedasticity". Journal of Econometrics. 31 : 307–327.

Box, G. E., Jenkins, G. M., Reinsel, G. C., and Ljung, G. M. (2015), Time series analysis: forecasting and control, John Wiley & Sons.

Chen, S., Chen, C. Y. H., Härdle, W. K., Lee, T. M., Ong, B. (2016), ``A first econometric analysis of the CRIX family." SFB 649 Discussion Paper forthcoming in Digital Banking and Internet Finance.

Engle, R. F. (1982). ``Autoregressive conditional heteroscedasticity with estimates of the variance of United Kingdom inflation." Econometrica: Journal of the Econometric Society, 987-1007.

Nielsen, F. A. (2011). ``A new ANEW: Evaluation of a word list for sentiment analysis in microblogs." In Proc. ESWC-11.

Nakamoto, S. (2008). ``Bitcoin: A peer-to-peer electronic cash system." Consulted, 1:2012, 2008.

Salton, G. McGill, M.(1983) editors. Introduction to Modern Information Retrieval. McGraw-Hill,

Schwarz, G. (1978), "Estimating the dimension of a model", Annals of Statistics, 6, 461–464.

Tang, X. (2005). ``Autoregressive hidden markov model with application in an El Nino study.", Master Thesis, University of Saskatchewan Saskatoon, 2004.

Rabiner, L. R. (1989).``A tutorial on hidden Markov models and selected applications in speech recognition." Proceedings of the IEEE, 77, 257-286.

中央銀行網站-發行貨幣-統計資料(https://www.cbc.gov.tw/np.asp?ctNode=409&mp=1)

幣虎(CoinGecko) (https://www.coingecko.com/en)

Bitcoin Forum (https://bitcointalk.org/index.php)

Poloniex Digital Asset Exchange(https://poloniex.com)