簡易檢索 / 詳目顯示
| 研究生： |
林子翔 Lin, Tzu Hsiang |
|---|---|
| 論文名稱： |
輿論對外匯趨勢的影響 The effects of public opinions on exchange rate movements |
| 指導教授： |
蔡瑞煌
Tsaih, Rua Huan |
| 口試委員: |
郭炳伸
Kuo, Biing Shen 蔡文禎 Tsay, Wen Jen |
| 學位類別： |
碩士
Master |
| 系所名稱： |
商學院 - 資訊管理學系 Department of Management Information System |
| 論文出版年： | 2017 |
| 畢業學年度： | 105 |
| 語文別： | 英文 |
| 論文頁數： | 63 |
| 中文關鍵詞： | 文字探勘 、機器學習 、匯率 、類神經網路 、TensorFlow 、圖形處理器 |
| 外文關鍵詞： | Text mining, Machine learning, Exchange rates, Artificial neural networks, Tensorflow, Graphic processing units |
| 相關次數： | 點閱：43 下載：13 |
| 分享至: |
| 查詢本校圖書館目錄 查詢臺灣博碩士論文知識加值系統 勘誤回報 |
本研究要探討的是在新聞、論壇和社群媒體討論的相關訊息是否真的會影響匯率的運動的假設。對於這樣的研究目標，我們建立了一個實驗，首先以文字探勘技術應用在新聞、論壇與社群媒體來產生與匯率相關的數值表示。接著，機器學習技術應用於學習得到的數值表示和匯率波動之間的關係。最後，我們證明透過檢驗所獲得的關係的有效性的假設。在此研究中，我們提出一種兩階段的神經網路來學習與預測每日美金兌台幣匯率的走勢。不同於其他專注於新聞或者社群媒體的研究，我們將他們進行整合，並將論壇的討論納為輸入資料。不同的資料組合產生出多種觀點，而三個資料來源的不同組合可能會以不同的方式影響預測準確率。透過該方法，初步實驗的結果顯示此方法優於隨機漫步模型。
This study wants to explore the hypothesis that the relevant information in the news, the posts in forums and discussions on the social media can really affect the daily movement of exchange rates. For such study objective, we set up an experiment, where the text mining technique is first applied to the news, the forum and the social media to generate numerical representations regarding the textual information relevant with the exchange rate. Then the machine learning technique is applied to learn the relationship between the derived numerical representations and the movement of exchange rates. At the end, we justify the hypothesis through examining the effectiveness of the obtained relationship. In this paper, we propose a hybrid neural networks to learn and forecast the daily movements of USD/TWD exchange rates. Different from other studies, which focus on news or social media, we integrate them and add the discussion of forum as input data. Different data combinations yield many views while different combination of three data sources might affect the forecasting accuracy rate in different ways. As a result of this method, the experiment result was better than random walk model.
Chapter 1. Introduction 1
1.1 Background 1
1.2 Motivation 3
1.3 Objective 6
Chapter 2. Literature Review 7
2.1 Background of Exchange Rates 7
2.2 Purchasing Power Parity (PPP) 8
2.3 Autoregressive Integrated Moving Average (ARIMA) Model 10
2.4 Random Walk Theory 12
2.5 Text Mining 14
2.6 Decision Support Mechanism 15
1. Concept Drifting 15
2. SLFN 16
3. The Resistant Learning with Envelope Module 17
4. Moving Window 19
2.7 Background of TensorFlow and GPU 20
1. TensorFlow 20
2. Graphics Processing Unit 23
3. TensorFlow and GPU implementation 24
2.8 Reasoning Neural Networks 24
Chapter 3. Experiment Design 26
3.1 Data Collection and Data Pre-Processing 26
1. Text Segmentation 27
2. Stop Words Removal 28
3. Part-of-Speech Tagging (PoS) 28
4. Sentiment Analysis 29
3.2 Build a Neural Network in TensorFlow 30
Chapter 4. Experimental Results 36
1. Facebook 39
2. News 41
3. Forum(PTT)： 44
4. Facebook and News: 46
5. Facebook and Forum (PTT) 49
6. News and Forum (PTT) 51
7. Facebook, News and Forum 54
Chapter 5. Conclusions and Future Works 57
5.1 Conclusions 57
5.2 Future Works 57
References 59
List of Tables
Table 1. The resistant learning with envelope module (Huang et al., 2014) 18
Table 2. Examples for TensorFlow operations (Abadi et al., 2016). 22
Table 3. Traditional Chinese Text Segmentation Example 28
Table 4. Example of Traditional Chinese stop words removal 28
Table 5. Example of Traditional Chinese part-of speech tagging 29
Table 6. Example of NTUSD words 29
Table 7. The definition of six variables in the input vector 31
Table 8. Representation for three movements groups 31
Table 9. Defintion of all variables 34
Table 10. Example of after transformed data 36
Table 11. Forecasting Accuracy Rate (Facebook) 40
Table 12. Forecasting Accuracy Rate (News) 43
Table 13. Forecasting Accuracy Rate (Forum) 45
Table 14. Forecasting Accuracy Rate (Facebook and News) 48
Table 15. Forecasting Accuracy Rate (Facebook and Forum) 50
Table 16. Forecasting Accuracy Rate (News and Forum) 53
Table 17. Accuracy of forecasting exchange rate movements (three sources) 55
List of Figures
Figure 1. Computation graph in TensorFlow (Abadi et al., 2016) 21
Figure 2. Network structure of RN 25
Figure 3. The flow chart of data collection and data pre-processing 27
Figure 4. Real exchange rate movements for USD/TWD 30
Figure 5. The flowchart of the learning process of Reasoning Neural Networks 32
Figure 6. The Thinking Mechanism 33
Figure 7. The Cramming Mechanism 33
Figure 8. The Reasoning Mechanism 34
Figure 9. The proposed neural network initial architecture 35
Figure 10. Volume of discussion and their positive / negative emotion 37
Figure 11. The implementation of moving windows in this experiment 38
Figure 12. Facebook Forecasting Result (M = 1~3) 39
Figure 13. Facebook Forecasting Result (M = 4~6) 40
Figure 14. News Forecasting Result (M = 1~3) 42
Figure15. News Forecasting Result (M = 4~6) 43
Figure 16. Forum Forecasting Result (M = 1~3) 44
Figure 17. Forum Forecasting Result (M = 4~6) 45
Figure 18. Facebook and News Forecasting Result (M = 1~3) 47
Figure 19. Facebook and News Forecasting Result (M = 4~6) 47
Figure 20. Facebook and Forum Forecasting Result (M = 1~3) 49
Figure 21. Facebook and Forum Forecasting Result (M = 4~6) 50
Figure 22. News and Forum Forecasting Result (M = 1~3) 52
Figure 23. News and Forum Forecasting Result (M = 4~6) 52
Figure 24. Facebook, News and Forum Forecasting Result (M = 1~3) 54
Figure 25. Facebook, News and Forum Forecasting Result (M = 4~6) 55
Abadi, M., Agarwal, A., Barham, P., Brevdo, E., Chen, Z., Citro, C., ... &Ghemawat, S. (2016). Tensorflow: Large-scale machine learning on heterogeneous distributed systems. arXiv preprint arXiv:1603.04467.
Abuaf, N., & Jorion, P. (1990). Purchasing power parity in the long run. The Journal of Finance, 45(1), pp. 157-174.
Babcock, B., Datar, M., & Motwani, R. (2002, January). Sampling from a moving window over streaming data. In Proceedings of the thirteenth annual ACM-SIAM symposium on Discrete algorithms (pp. 633-634). Society for Industrial and Applied Mathematics.
Babaeizadeh, M., Frosio, I., Tyree, S., Clemons, J., Kautz, J. (2017). Reinforcement Learning through Asynchronous Advantage Actor-Critic on a GPU. ICLR 2017 conference submission
Barzdins, G., Renals, S., & Gosko, D. (2016). Character-Level Neural Translation for Multilingual Media Monitoring in the SUMMA Project. arXiv preprint arXiv:1604.01221.
Elwell, R. &. (2011). Incremental learning of concept drift in nonstationary environments. IEEE Transactions on Neural Networks, 22(10), pp. 1517-1531.
Fama, E. F. (1965). The behavior of stock-market prices. The journal of Business, 38(1), pp. 34-105.
Fama, E. F. (1970). Efficient capital markets: A review of theory and empirical work. The journal of Finance, 25(2), pp. 383-417.
Feldman, R., & Sanger, J. (2007). The text mining handbook: advanced approaches in analyzing unstructured data. Cambridge university press.
Frenkel, J. A. (1981). The collapse of purchasing power parities during the 1970's. European Economic Review, 16(1), pp. 145-165.
G.E.P. Box, G. J. (1970). Time series analysis, forecasting and control. Holden-Day, San Francisco, CA.
Gama, J., Žliobaitė, I., Bifet, A., Pechenizkiy, M., &
Bouchachia, A. (2014). A survey on concept drift adaptation. ACM Computing Surveys (CSUR), 46(4), p. 44.
Hendranata, A. (2003). ARIMA (Autoregressive Moving Average). Manajemen Keuangan Sektor Publik FEUI.
Hogan, L. I. (1986). A comparison of alternative exchange rate forecasting models. Economic Record, 62(2), pp. 215-223.
John Walker, S. (2014). Big data: A revolution that will transform how we live, work, and think.
Kilburn, F. (2014, Mar 4). Markit Taps SMA for Social Media Sentiment Indicators. Retrieved from waterstechnology: http://www.waterstechnology.com/inside-market-data/news/2332009/markit-taps-sma-for-social-media-sentiment-indicators
Levinson, M. (2014). The Economist Guide to Financial Markets: Why they exist and how they work. PublicAffairs.
Lun-Wei Ku and Hsin-Hsi Chen (2007). Mining Opinions from the Web: Beyond Relevance Retrieval. Journal of American Society for Information Science and Technology, Special Issue on Mining Web Resources for Enhancing Information Retrieval, 58(12), pages 1838-1850.
MacDonald, R., & Marsh, I. W. (1997). On fundamentals and exchange rates: a Casselian perspective. Review of Economics and Statistics, 79(4), pp. 655-664.
Madura, J. (2011). International financial management. Cengage Learning.
Mankiw, N. G. (2010). Macroeconomics. Cengage Learning.
Masud, M. G. (2011). Classification and novel class detection in concept-drifting data streams under time constraints. IEEE Transactions on Knowledge and Data Engineering, 23(6), pp. 859-874.
Meese, R. A. (1983). Empirical exchange rate models of the seventies: Do they fit out of sample?. Journal of international economics, 14(1-2). pp. 3-24.
Meyler, A., Kenny, G., & Quinn, T. (1998). Forecasting Irish inflation using ARIMA models.
NVidia, F. (2009). Nvidia’s next generation cuda compute architecture. NVidia, Santa Clara, Calif, USA.
Osborne, J. (2016, 8 22). Google's Tensor Processing Unit explained: this is what the future of computing looks like. Retrieved from techradar: http://www.techradar.com/news/computing-components/processors/google-s-tensor-processing-unit-explained-this-is-what-the-future-of-computing-looks-like-1326915
Owens, J. D., Houston, M., Luebke, D., Green, S., Stone, J. E., & Phillips, J. C. (2008). GPU computing. Proceedings of the IEEE, 96(5), 879-899.
Pai, P. F., & Lin, C. S. (2005). A hybrid ARIMA and support vector machines model in stock price forecasting. Omega, 33(6), pp. 497-505.
Pearson, K. (1905). The problem of the random walk. . Nature, 72(1865), 294.
Peramunetilleke, D., & Wong, R. K. (2002). Currency exchange rate forecasting from news headlines. Australian Computer Science Communications, 24(2), pp. 131-139.
Roberts, H. V. (1959). Stock‐Market “Patterns” And Financial Analysis: Methodological Suggestions. The Journal of Finance, 14(1), pp. 1-10.
Sabur, S. A., & Molla, A. R. (1993). Trend, Variability and Relative Profitability of spices in Bangladesh. Bangladesh Journal of Agricultural Economics, 316(1).
Salvatore, D. (2012). International economics. Wiley Global Education.
Seelenfreund, A., Parker, G. G., & Van Horne, J. C. (1968). Stock price behavior and trading. journal of Financial and Quantitative Analysis, 3(03),, pp. 263-281.
Sekine, M. T. (2001). Modeling and forecasting inflation in Japan (No. 1-82). International Monetary Fund.
Stanley, K. O. (2003). Learning concept drift with a committee of decision trees. Informe técnico: UT-AI-TR-03-302.
Szegedy, C., Ioffe, S., Vanhoucke, V., & Alemi, A. (2016). Inception-v4, inception-resnet and the impact of residual connections on learning. arXiv preprint arXiv:1602.07261.
Tsaih, R. H., & Cheng, T. C. (2009). (2009). A resistant learning procedure for coping with outliers. Annals of Mathematics and Artificial Intelligence, 57(2), pp. 161-180.
Tsaih, R. R. (1998). An explanation of reasoning neural networks. Mathematical and computer modelling, 28(2), pp. 37-44.
Tsymbal, A. (2004). The problem of concept drift: definitions and related work. Computer Science Department, Trinity College Dublin, p. 106(2). Retrieved from Computer Science Department, Trinity College Dublin: 106(2)
Varian, H. R. (2014). Big data: New tricks for econometrics. . The Journal of Economic Perspectives, 28(2), pp. 3-27.
Wood, D., & Dasgupta, B. (1996). Classifying trend movements in the MSCI USA capital market index—a comparison of regression, ARIMA and neural network methods. Computers & Operations Research, 23(6), pp. 611-622.
Zhang, D., Simoff, S., & Debenham, J. (2005). Exchange rate modelling using news articles and economic data . AI 2005: Advances in Artificial Intelligence, pp. 467-476.
Zhang, G. P. (1998). Forecasting with artificial neural networks: The state of the art. . International journal of forecasting, 14(1), pp. 35-62.
Zhang, G. P. (2003). Time series forecasting using a hybrid ARIMA and neural network model. . Neurocomputing 50, pp. 159-175.
