(3.238.174.50) 您好!臺灣時間:2021/04/18 02:37
字體大小: 字級放大   字級縮小   預設字形  
回查詢結果

詳目顯示:::

我願授權國圖
: 
twitterline
研究生:廖亮瑋
研究生(外文):Liang-Wei Liao
論文名稱:使用長短期記憶遞歸神經網路之匯率預測模型—考慮財經變數與財經新聞
論文名稱(外文):Exchange Rate Forecasting using Long Short Term Memory Networks — Considering Economic Variables and Financial News
指導教授:呂永和
指導教授(外文):Yung-Ho Leu
口試委員:楊維寧陳雲岫
口試委員(外文):Wei-Ning YangYun-Shiow Chen
口試日期:2018-07-22
學位類別:碩士
校院名稱:國立臺灣科技大學
系所名稱:資訊管理系
學門:電算機學門
學類:電算機一般學類
論文種類:學術論文
論文出版年:2019
畢業學年度:107
語文別:中文
論文頁數:64
中文關鍵詞:深度學習長短期記憶遞歸神經網路情緒分析匯率預測
外文關鍵詞:Deep LearningLong Short-Term Memory Recurrent Neural NetworkSentiment AnalysisExchange Rate Forecasting
相關次數:
  • 被引用被引用:2
  • 點閱點閱:259
  • 評分評分:系統版面圖檔系統版面圖檔系統版面圖檔系統版面圖檔系統版面圖檔
  • 下載下載:0
  • 收藏至我的研究室書目清單書目收藏:0
固定匯率制度的瓦解,象徵政府操作匯率的干預力道減弱,經貿市場走向全面自由化,台灣也乘著這股浪潮,在國際貿易中扮演不可或缺的角色。然而,頻繁的經貿活動意味著投資買賣風險增加,而匯率變動體現的是一國貨幣購買力出現變化,若無法掌握匯率的升貶,做出錯誤的投資買賣決策,將造成國家企業的巨大損失。因此,預測匯率是許多財經專家與研究人員所關注的議題。
過往針對匯率的研究中,多是探討匯率與各項潛在因子的互動關係,或是運用基於經濟學理論所提出的線性迴歸模型預測匯率。然而,上述兩種研究方法多著墨於總體經濟變數與技術指標的探討,較少利用新聞文本作為預測匯率之依據,因此,本研究參考情緒詞典BosonNLP,考量新聞文本內的情緒詞彙與情緒組合詞以進行情緒分析,藉此表達其對於未來匯率漲跌之信心程度。此外,本研究也彙整過往相關研究所探討的匯率影響因子並進行變數篩選,找出與匯率漲跌最直接相關的變數。最後,採用可以很好地處理時間序列的長短期記憶遞歸神經網路(Long Shot Term Memory Recurrent Neural Network, LSTM)進行匯率預測。實驗結果發現,基於經濟學理論的隨機漫步模型(Random Walk, RW)與廣義自迴歸條件異方差模型(Generalized Autoregressive Conditional Heteroskedasticity, GARCH)的均方根誤差(Root Mean Square Error, RMSE)與預測漲跌準確率(Direction Accuracy, DA)分別為0.119及46.44%與0.083及50.76%,而長短期記憶遞歸神經網路則分別為0.079與64.21%,而在加入新聞文本輔以預測後,RMSE及DA分別進步至0.077與68.27%,尤其在預測匯率變動的方向準確率上得到4.07%的明顯提升。
With the collapse of fixed exchange rate system, the power of the government intervention to control the exchange rate diminished and the financial market has become a free market. Taiwan also plays an important role in the liberalization of international trade. Since frequent trading activities and the investment risk usually come hand in hand, the government and the enterprises will suffer severe capital loss if they are not able to forecast exchange rate so as to make a bad investment decision. Therefore, exchange rate forecasting has become an important research issue for financial experts and researchers.
Most of the previous researches either explored the cause-effect relation between exchange rate and potential factors or forecast exchange rate using linear regression model. Not many researches today have explored enough the effect of textual news on exchange rate. In this thesis, we explore the effect of news articles on exchange rate based on sentiment analysis. To measure the effect of news articles on exchange rate, we extracted sentiment-based words and sentiment-based compound words to calculate the sentimental strength of news articles based on the sentiment lexicon. Besides, to construct an exchange rate forecasting model, we select prediction variables which are the most relevant to the exchange rate forecasting. With the sentiment-based words and the selected variables, we constructed a model to predict the exchange rate fluctuations using the long short-term memory recurrent neural network model. The experiment result showed that the RMSE and the direction accuracy of the random walk model are 0.119 and 46.44%, respectively, while those of the GARCH model are 0.083 and 50.76%, respectively. For the LSTM model, the RMSE and the directional accuracy are 0.079 and 64.21%, respectively. In comparison, the proposed LSTM model which considered the news articles as an additional factor showed a significant improvement on directional accuracy over the pure LSTM model by 4.07%.
摘要 I
ABSTRACT II
誌謝 III
目錄 IV
圖目錄 VI
表目錄 VII
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 2
1.4 研究架構 3
第二章 文獻回顧 5
2.1匯率與財經變數的關聯性 5
2.1.1 匯率與總體經濟變數 5
2.1.2 匯率與技術指標 7
2.2 深度學習相關模型 9
2.2.1 遞歸神經網路模型 9
2.2.2 Word2Vec模型 20
2.3 神經網路預測匯率相關研究 30
2.4 情緒分析相關研究 32
第三章 研究方法 35
3.1 研究方法架構 35
3.2 資料來源與變數篩選 36
3.2.1 潛在預測變數資料來源 36
3.2.2 變數篩選 37
3.3 中文斷詞 37
3.4 文本情緒分析 38
3.5 預測模型選取 40
3.6 預測模型比較 41
3.7 預測模型衡量指標 42
第四章 實驗結果與分析 44
4.1 變數篩選結果 44
4.2 模型比較結果 45
4.3 加入新聞文本比較結果 46
第五章 結論與未來展望 49
5.1 結論 49
5.2 未來展望 49
參考文獻 50
[1] C. Bonser-Neal, V.V. Roley, and G. H. Sellon, Jr., “Monetary Policy Actions,
Intervention, and Exchange Rates: A Reexamination of the Empirical
Relationships Using Federal Funds Rate Target Data,” The Journal of Business,
University of Chicago Press, vol. 71(2), pp. 147-177, April 1998.
[2] Ž. Živković, I. Mihajlović, and Đ. Nikolić, “Artificial Neural Network Method
Applied on the Nonlinear Multivariate Problems,” Serbian Journal of
Management, vol. 4, no. 2, pp. 143-155, 2009.
[3] A. El-Shafie and A. Noureldin, “Generalized versus Non-generalized Neural
Network Model for Multi-lead Inflow Forecasting at Aswan High Dam,”
Hydrology and Earth System Sciences, vol. 7, no. 5, pp. 7957-7993, Oct. 2011.
[4] A. Landi, P. Piaggi, M. Laurino, and D. Menicucci, “Artificial Neural Networks
for Nonlinear Regression and Classification,” in Proc. 10th International
Conference on Intelligent Systems Design and Applications, Cairo, Egypt, pp. 115-
120, Nov. 2010.
[5] A. Peguin-Feissolle, “A Comparison of the Power of Some Tests for Conditional
Heteroscedasticity,” Economics Letters, vol. 63, pp. 5-17, May 1999.
[6] J. Elman, “Finding Structure in Time,” Cognitive Science, vol. 14, pp. 179-211,
1990.
[7] S. Hochreiter and J. Schmidhuber, “Long Short-Term Memory,” Neural
Computation, vol. 9, no. 8, pp. 1735-1780, 1997.
[8] T. Mikolov, K. Chen, G. Corrado, and J. Dean, “Efficient Estimation of Word
Representations in Vector Space,” in Proc. 1st International Conference on
Learning Representations, Sep. 2013.
[9] X. Rong, “word2vec Parameter Learning Explained,” arXiv preprint
51
arXiv:1411.2738, 2014.
[10] R. A. Meese and K. Rogoff, “Empirical Exchange Rate Models of the Seventies:
Do They Fit Out of Sample?,” Journal of International Economics, vol. 14, no. 1-
2, pp. 3-24, February 1983.
[11] L. Chin, M. Azali, Z. B. Yusop, and M. B. Yusoff, “The Monetary Model of
Exchange Rate: Evidence from the Philippines,” Applied Economics Letters, vol.
14, pp. 993-997, Oct. 2007.
[12] M. D. Chinn, “Some Linear and Nonlinear Thoughts on Exchange Rates,”
Journal of International Money and Finance, vol. 10, pp. 214-230, 1991.
[13] C. Brooks, “Testing for Non-Linear in Daily Sterling Exchange Rates,” Applied
Financial Economics, vol. 6, no. 4, pp. 307-317, 1996.
[14] G. Boero and E. Marrocu, “The Performance of Non-linear Exchange Rate
Models: a Forecasting Comparison,” Journal of Forecasting, vol. 21, pp. 513-542,
July 2002.
[15] C. Panda and V. Narasimhan, “Forecasting Exchange Rate Better with Artificial
Neural Network,” Journal of Policy Modeling, vol. 29, pp. 227-236, 2007.
[16] J. Yao and C. L. Tan, “A Case Study on Using Neural Networks to Perform
Technical Forecasting of Forex,” Neurocomputing, vol. 34, pp. 79-98, April 2000.
[17] Z. C. Lipton, J. Berkowitz, and C. Elkan, “A Critical Review of Recurrent Neural
Networks for Sequence Learning,” Computing Research Repository,
arXiv:1506.00019v4, Oct. 2015.
[18] S. S. Saini, O. Parkhe, and T. D. Khadtare, “Analysis of Feedforward and
Recurrent Neural Network in Forecasting Foreign Exchange Rate,” Imperial
Journal of Interdisciplinary Research, vol. 2, no. 6, pp. 822-826, 2016.
[19] P. Tenti, “Forecasting Foreign Exchange Rates Using Recurrent Neural
52
Networks,” Applied Artificial Intelligence, vol. 10, pp. 567-581, 1996.
[20] Doaa, Mohey, El-Din, Mohamed, and Hussein, “A Survey on Sentiment Analysis
Challenges,” Journal of King Saud University – Engineering Sciences, vol. 34, no.
4, April 2016.
[21] S. M. Vohra and J. B. Teraiya, “A Comparison Study of Sentiment Analysis
Techniques,” Journal of Information, Knowledge and Research in Computer
Engineering, vol. 2, no. 2, pp. 313-317, Oct. 2013.
[22] D. Tang, F. Wei, B. Qin, N. Yang, T. Liu, and M. Zhou, “Sentiment Embeddings
with Applications to Sentiment Analysis,” IEEE Transactions on Knowledge and
Data Engineering, vol. 28, no. 2, pp. 496-509, February 2016.
[23] L. C. Yu, J. Wang, K. R. Lai, and X. Zhang, “Refining Word Embeddings for
Sentiment Analysis,” in Proc. 2017 Conference on Empirical Methods in Natural
Language Processing, pp. 534-539, Sep. 2017.
[24] S. Kiritchenko and S. M. Mohammad, “The Effect of Negators, Modals, and
Degree Adverbs on Sentiment Composition,” in Proc. NAACL-HLT 2016, pp. 43-
52, June 2016.
[25] Z. Yangsen, J. Yuru, and T. Yixuan, “Study of Sentiment Classification for
Chinese Microblog Based on Recurrent Neural Network,” Chinese Journal of
Electronics, vol. 25, no. 4, pp. 601-607, July 2016.
[26] 楊琬竹(2016),台灣加權股價指數、美元匯率及杜拜油價之關聯性研究,碩士論文,國立高雄應用科技大學。
[27] 王秋懿(2016),台灣加權股價指數、美國道瓊工業指數、中國A 股與匯率之研究,碩士論文,國立高雄應用科技大學。
[28] 林杏津(2016),探討匯率與台灣加權股價指數之相互影響—以美元、歐元、人民幣、日圓、港幣為例,碩士論文,亞洲大學。
[29] 施佳恩(2015),台灣加權指數、匯率、那斯達克指數之關聯性研究,碩士論文,國立高雄應用科技大學。
[30] 張志賓(2011),台灣加權指數、上海指數、香港指數、美元匯率、石油、黃金、道瓊指數關聯性分析,碩士論文,國立雲林科技大學。
[31] 羅碧霞(2012),比較美元指數與新台幣匯率對台灣股價之互動關係,碩士論文,淡江大學。
[32] 蔡依靜(2017),油價、新台幣匯率與商品價格之關聯性研究,碩士論文,國立屏東科技大學。
[33] 陳谷劦(2003),公債殖利率與總體變數間關係之探討-台灣的實證研究,碩士論文,世新大學。
[34] 林佩君(2016),台灣、美國兩國間利率、股價、貨幣供給與匯率的關聯性,碩士論文,亞洲大學。
[35] 巴家駿(2016),匯率波動下對股市報酬及利率的影響,碩士論文,東吳大學。
[36] 黃素琴(2013),技術分析運用在外匯市場之有效性探討-以新台幣及世界主要六國貨幣兌美元之匯率為例,碩士論文,明新科技大學。
[37] 張淵植(2008),應用技術分析探討外匯市場效率性,碩士論文,國立東華大學。
[38] 賈格賽漢(2016),比較技術指標之MA、MACD 及RSI 應用於外匯市場之操作績效,碩士論文,南華大學。
[39] 王界舜(2010),外匯技術指標獲利績效分析與預測,碩士論文,輔仁大學。
[40] 蔡宗岸(2007),簡單技術分析交易法則,碩士論文,國立政治大學。
[41] 陳有忠(2007),技術指標應用於外匯交易之研究-以英鎊、日圓及台幣為例,碩士論文,國立臺灣科技大學。
[42] 顏伯欣(2017),應用技術指標於外匯市場自動交易實證之研究,碩士論文,國立勤益科技大學。
[43] 王侑琳(2012),運用技術分析探討廠商外匯資產配置-以新幣與韓圜為例,碩士論文,國立臺灣大學。
[44] 張天禹(2015),技術分析應用於外匯交易之研究,碩士論文,東吳大學。
[45] 郭峻志(2017),外匯程式交易策略評估之實證研究,碩士論文,國立勤益科技大學。
連結至畢業學校之論文網頁點我開啟連結
註: 此連結為研究生畢業學校所提供,不一定有電子全文可供下載,若連結有誤,請點選上方之〝勘誤回報〞功能,我們會盡快修正,謝謝!
QRCODE
 
 
 
 
 
                                                                                                                                                                                                                                                                                                                                                                                                               
第一頁 上一頁 下一頁 最後一頁 top
系統版面圖檔 系統版面圖檔