臺灣博碩士論文加值系統

配對交易 (pair trading) 又稱為價差交易。此交易策略運用穩定 (stationary) 時間序列具有均數回歸 (mean reversion) 的性質，將市場上兩檔具有高度相關股價走勢的股票形成一組配對股票。當此配對股票之投資組合價值偏離歷史均值，將藉由一多一空的組合交易進行套利 (arbitrage)。本論文使用台灣經濟新報 (TEJ) 提供之2016年台灣上市股票之日調整收盤價資料，藉由 Said and Dickey (1984) 提出修正之 augmented Dickey-Fuller 檢定測試一個自回歸模型 (autoregression model) 是否存在單根 (unit root)，而由 Engle and Granger (1987) 的共整合檢定 (cointegration test) 測試任兩檔股票是否具有共整合性質，而考量成為配對交易之股票。由此得到大量組數的配對股票，形成一個大型的複雜網絡 (network) 系統。實際應用上，配對交易股票所形成的網絡圖可能更複雜且數據量更龐大。由於 Spark 有效提供了處理複雜圖形之演算法的工具，因此本文以 Spark 之社群網絡分析 (social network analysis) 系統 GraphFrames 為工具，採用有效且能快速取得高品質分群結果的標籤傳播演算法 (label propagation algorithm)，對此大型複雜網絡進行社群發現 (community detection)。最後由配對股票的分群關係，進一步分析2016年的分群結果與個股之間的關聯。

Pair trading, also called as Spreading, is a trading strategy which can make two stocks whose prices are on a high correlation to be a pair based on the property of mean reversion in a stationary time series. Furthermore, it can be arbitraged through a portfolio with a long position and a short position as the value of the portfolio deviates from the historical mean. The thesis will apply the daily adjusted closing prices of publicly traded companies in 2016 from TEJ to test if an autoregressive model exists a unit root or not by augmented Dickey-Fuller Test of Said and Dickey (1984), and then test if any two stocks exist the property of cointegration by Cointegration Test of Engle and Granger (1987). If two stocks exist the property of cointegration, they will be assumed to be a pair. There are lots of pairs constitute a huge and complicated network system in the thesis. In fact, there are many effective algorithms in Spark to deal with complicated graphs, and that is why the thesis applies its tools, GraphFrames, to analyze the complex network by Social Network Analysis. Eventually, the thesis will use Label Propagation Algorithm to detect communities in the network, and then work on the further analysis of the result.

摘要 I
Abstract II
目錄 III
圖目錄 IV
表目錄 V
1. 前言 1
2. 配對交易與共整合檢定 2
3. 社群網絡分析與 Spark 之分析工具 GraphFrames 5
4. 實證結果 16
5. 結論 34
未來研究方向 35
參考文獻 35

余佳倩 (2015)，“以台灣金融業股票探討流動性風險與資產定價之關係” 碩士論文，台中：國立中興大學財務金融研究所。
周群哲 (2016)，“全球股市流動性之關連性分析” 碩士論文，桃園：國立中央大學經濟學研究所。
蘇怡仁，李哲均，張傳旺，李俊德 (2015)，”以超圖分割法實現社群發現之研究” 第十四屆離島資訊技術與應用研討會，澎湖：國立澎湖科技大學資訊工程系。
Amihud, Y., (2002). Illiquidity and stock returns: cross-section and time-series effects. Journal of Financial Markets, 5(1), 31-56.
Amr, T. and R. Stamboliyska, (2016). Practical D3.js: Master the Use of D3.js in the Real World. Apress, New York.
Bedi, P. and C. Sharma, (2016). Community detection in social networks. WIREs Data Mining and Knowledge Discovery, 6(3), 115-135.
Beeley, C., (2016). Web Application Development with R Using Shiny. Second Edition. Packt Publishing, UK.
Davey, B.A. and J. Pitkethly, (2013). Google pagerank. Australian Mathematical Sciences Institute, Australia.
Dickey, D.A. and W.A. Fuller, (1979). Distribution of the estimators for autoregressive time series with a unit root. Journal of the American Statistical Association, 74(366), 427-431.
Engle, R.F. and C.W.J. Granger, (1987). Cointegration and error correction, estimation, and testing. Econometrica, 55(2), 251-276.
Florackis, C., A. Gregoriou and A. Kostakis (2011). Trading frequency and asset pricing on the London Stock Exchange: evidence from a new price impact ratio. Journal of Banking & Finance, 35(2), 3335-3350.
Granger, C.W.J. and P. Newbold, (1974). Spurious regressions in econometrics. Journal of Econometrics, 2(2), 111-120.
Guller, M., (2016). Big Data Analytics with Spark: A Practitioner’s Guide to Using Spark for Large Scale Data Analysis. Apress, New York.
Iteydt, M., (2015). D3.js by Example: Create Attractive Web-Based Data Visualizations Using the Amazing Javascript Library D3.js. Packt Publishing, UK.
Lancichinetti, A. and S. Fortunato, (2009). Community detection algorithms: A comparative analysis. Physical Review E, 80(5), 056117. Erratum: Physical Review E, 89(4), 049902 (2014).
LazyProgrammer, (2016). Big Data, MapReduce, Hadoop, and Spark with Python. http://lazyprogrammer.me.
Liu, B. and P.S. Yu, (2009). PageRank. In The Top Ten Algorithms in Data Mining. Edited by X. Wu and V. Kumar, 117-125. CRC Press, New York.
Meeks, E., (2015). D3.js in Action. Manning Publications Co., New York.
Murray, S., (2013). Interactive Data Visualization for the Web. O’Reilly Media, Inc., Beijing.
Nelson, C. and C. Plosser (1982). Trends and random walks in macroeconomic time series: Some evidence and implications. Journal of Monetary Economics, 10(2), 139-162.
Newman, M.E.J. and M. Girvan, (2004). Finding and evaluating community structure in networks. Physical Review E, 69(2), 026113.
Page, L., S. Brin, R. Motwani, and T. Winograd, (1998). The PageRank citation ranking: bringing order to the web. Proceedings of the 7th International World Wide Web Conference, 161-172. Brisbane, Australia.
Raghavan, U.N., R. Albert, and S. Kumara, (2007). Near linear time algorithm to detect community structures in large-scale network. Physical Review E, 76(3), 036106.
Resnizky, H.G., (2015). Learning Shiny. Packt Publishing, UK.
Said, S.E. and D.A. Dickey, (1984). Test for unit roots in autoregressive-moving average models of unknown order. Biometrika, 71(3), 599-607.
Sankar, K., (2016). Fast Data Processing with Spark 2. Third Edition. Packet Publishing, UK.
Traag, V., (2014). Algorithms and Dynamical Models for Communities and Reputation in Social Networks. Springer International Publishing AG, Switzerland.
Tsvetovat, M. and A. Kouznetsov, (2011). Social Network Analysis for Startups. O’Reilly Media, Inc., Beijing.
Zafarani, R., A. Abbasi and H. Liu, (2014). Social Media Mining: An Introduction. Cambridge University Press, UK.