以深度學習為基礎解析美聯準會議紀要 對美元指數之預測研究

以作者查詢圖書館館藏

、以作者查詢臺灣博碩士

、以作者查詢全國書目

、勘誤回報

、線上人數：134

、訪客IP：3.133.79.70

姓名

葉建宏(Chien-Hung Yeh) 查詢紙本館藏

畢業系所

資訊管理學系在職專班

論文名稱

以深度學習為基礎解析美聯準會議紀要對美元指數之預測研究

相關論文

★ 台灣50走勢分析：以多重長短期記憶模型架構為基礎之預測	★ 以多重遞迴歸神經網路模型為基礎之黃金價格預測分析
★ 增量學習用於工業4.0瑕疵檢測	★ 遞回歸神經網路於電腦零組件銷售價格預測之研究
★ 長短期記憶神經網路於釣魚網站預測之研究	★ 基於深度學習辨識跳頻信號之研究
★ Opinion Leader Discovery in Dynamic Social Networks	★ 深度學習模型於工業4.0之機台虛擬量測應用
★ A Novel NMF-Based Movie Recommendation with Time Decay	★ 以類別為基礎sequence-to-sequence模型之POI旅遊行程推薦
★ A DQN-Based Reinforcement Learning Model for Neural Network Architecture Search	★ Neural Network Architecture Optimization Based on Virtual Reward Reinforcement Learning
★ 生成式對抗網路架構搜尋	★ 以漸進式基因演算法實現神經網路架構搜尋最佳化
★ Enhanced Model Agnostic Meta Learning with Meta Gradient Memory	★ 遞迴類神經網路結合先期工業廢水指標之股價預測研究

檔案

[Endnote RIS 格式]

[Bibtex 格式]

[相關文章]

[文章引用]

[完整記錄]

[館藏目錄]

至系統瀏覽論文 (2026-8-1以後開放)

摘要(中)

在全球化的時代，各國經濟依賴貿易的程度日漸提升，貨幣交換的匯率波動牽動著企業的利潤多寡、國家整體經濟的成長率，而美元為全球主要交易貨幣，美元匯率的趨勢對於全球各國經濟、企業或個人資產配置影響甚鉅，美國聯邦準備系統（Federal Reserve System，Fed）相當於各國央行的角色，簡稱美聯準，其主要使命包含擴大就業、穩定物價、調控長期利率等，其中以聯邦公開巿場操作委員會(Federal Open Market Committee, FOMC)為主要管理政策單位，每年召開8次會議並發布會議紀要，又稱為美聯準會議紀要，本研究目的在以美聯準會議紀要內容探求美聯準對於美國經濟的看法及各項公布數據對於美元指數的影響，以建構深度學習中的長短期記憶模型，判斷會議紀要全文內容中對美國整體經濟的看法或是採取的政策對於美元指數(US Dollar Index, USDX)的關聯。本研究在文字探勘、情緒分析的基礎上，萃取美聯準會議紀要中對於經濟有影響的重要的關鍵字詞，結合美元指數的歷史收盤價用以建構之FOMC-Explor模型，對美聯準會議紀要公布後30天的美元指數趨勢做出預測，對比單純以歷史匯價建構的長短期記憶模型(Long Short-Term Memory, LSTM)或單純以文本分析建構的長短期本研究建構的長短期記憶模型， FOMC-Explor模型對於美聯準會議紀要發布內容後對於美元指數造成的影響，能提供具有較穩定且誤差最小的美元指數趨勢預測，作為企業及個人避險與資產配置的參考依據。

摘要(英)

In the era of globalization, the economy of countries dependent on international trade. Currency is the medium of international trade and volatility of the exchange rate affects the profitability of enterprises and the growth rate of the country’s overall economy. U.S. dollar is the most popular valuation currency in international trade contract. The trend of the US dollar exchange rate has a huge impact on the global economy, corporate or personal asset allocation. The Federal Reserve System (Fed) play the role as central banks in United States of America. The missions of Fed are to expand employment, stabilize the inflation and regulate the long-term interest rates, etc. The Federal Open Market Committee (FOMC) is in charge of the monetary policy and holds 8 meetings each year and releases their meeting minutes, also known as the Federal Open Market Committee minutes. The purpose of this research is to explore the Fed’s expectation of the overall economic status from the published economic indexes and FOMC meeting minutes. Based on the text mining and sentiment analysis, this research extracts the keywords from the FOMC meeting minutes and then combines historical closed US dollar index (USDX) to construct the FOMC-Explore model. To compare the USDX trend prediction result between FOMC minutes released date to one month later then, FOMC-Explore model can provide a more stable prediction result than the Long Short-Term Memory (LSTM) model which use historical USDX or FOMC minutes. FOMC-Explore model prediction can provide corporate and individual investor a valuable reference to reduce the risk from volatility of the exchange rate and asset allocation.

關鍵字(中)

★ 長短期記憶模型
★ 美聯準會議紀要
★ 美元指數
★ FOMC-Explore模型

關鍵字(英)

★ Long Short-Term Memory
★ FOMC Meeting Minutes
★ USDX
★ FOMC-Explore Model

論文目次

中文摘要 I
Abstract II
誌謝 III
圖目錄 VI
第一章緒論 1
1.1 研究背景 1
1.2 研究目的 2
1.3 研究貢獻 3
1.4 論文架構 4
第二章文獻探討 5
2.1 美聯準會議紀要與美元匯率及金融市場相關研究 5
2.2 類神經網路預測價格相關研究 7
第三章研究方法 10
3.1 研究流程 10
3.2 資料處理 10
3.2.1 資料蒐集與前處理 10
3.3 模型建構 12
3.3.1 詞頻與逆向檔案頻率(Term Frequency-Inverse Document Frequency, TF-IDF) 13
3.3.2 長短期記憶模型(Long Short-Term Memory, LSTM) 14
3.4 模型預測 15
第四章實驗結果 16
4.1 實驗環境 16
4.2 資料集 17
4.3 實驗介紹: 17
4.3.1 實驗A: USDX-Based LSTM 17
4.3.2 實驗B: Article-Based LSTM 18
4.3.3 實驗C FOMC-Explore Model 19
4.4 參數比較 20
4.4.1 神經網路訓練周期數(Epoch)比較 21
4.4.2 學習率(Learning Rate)比較. 22
4.4.3 批次尺寸(Batch Size)比較 24
4.4.4 Dropout參數比較 25
4.5 模型驗證 27
4.5.1 均方誤差(Mean Squared Error, MSE) 27
4.5.2 實驗A: USDX-Based LSTM 28
4.5.3 實驗B: Article-Based 28
4.5.4 實驗C: FOMC-Explore 29
4.6 使用者案例 30
第五章結論 31
5.1 研究結果: 31
5.2 未來研究方向與建議 32
第六章參考資料 33

參考文獻

黃裕烈, & 管中閔. (2019). 美國聯準會會議紀要的文字探勘與台灣經濟變數預測. 經濟論文叢刊, 47(3), 363–391. https://doi.org/10.6277/TER.201909_47(3).0002
揪莉森. (2021, 七月 22). [精進魔法] Optimization：優化深度學習模型的技巧（中）－ Adaptive Learning Rates. iT 邦幫忙::一起幫忙解決難題，拯救 IT 人的一天. https://ithelp.ithome.com.tw/articles/10204032
深度学习. (2021). 收入維基百科，自由的百科全書. https://zh.wikipedia.org/w/index.php?title=%E6%B7%B1%E5%BA%A6%E5%AD%A6%E4%B9%A0&oldid=64709501
神經網路中Epoch、Iteration、Batchsize相關理解和說明 | 程式前沿. (不詳). 讀取於 2021年7月31日, 從 https://codertw.com/程式語言/557816/
王姿華. (2020). 利用階層式注意力深度學習模型從金融新聞預測匯率波動 A hierarchical attention model for volatility of the exchange rate prediction using financial news [東吳大學巨量資料管理學院]. https://hdl.handle.net/11296/dgr5e3
Boukus, E., & Rosenberg, J. V. (2006). The Information Content of FOMC Minutes (SSRN Scholarly Paper ID 922312). Social Science Research Network. https://doi.org/10.2139/ssrn.922312
Feldman, R., & Sanger, J. (2007). The Text Mining Handbook: Advanced Approaches in Analyzing Unstructured Data. Cambridge University Press.
Huang, R. D., & Masulis, R. W. (1999). FX Spreads and Dealer Competition Across the 24-Hour Trading Day. The Review of Financial Studies, 12(1), 61–93. https://doi.org/10.1093/rfs/12.1.61
Huang, Y.-L., & Kuan, C.-M. (2021). Economic prediction with the FOMC minutes: An application of text mining. International Review of Economics & Finance, 71, 751–761. https://doi.org/10.1016/j.iref.2020.09.020
Jung, G., & Choi, S.-Y. (2021). Forecasting Foreign Exchange Volatility Using Deep Learning Autoencoder-LSTM Techniques. Complexity, 2021, e6647534. https://doi.org/10.1155/2021/6647534
Kim, S. (2020). Macroeconomic and Financial Market Analyses and Predictions Through Deep Learning (SSRN Scholarly Paper ID 3684936). Social Science Research Network. https://doi.org/10.2139/ssrn.3684936
Livieris, I. E., Pintelas, E., & Pintelas, P. (2020). A CNN–LSTM model for gold price time-series forecasting. Neural Computing and Applications, 32(23), 17351–17360. https://doi.org/10.1007/s00521-020-04867-x
Maneejuk, P., & Srichaikul, W. (2021). Forecasting foreign exchange markets: Further evidence using machine learning models. Soft Computing, 25(12), 7887–7898. https://doi.org/10.1007/s00500-021-05830-1
Maqsood, H., Mehmood, I., Maqsood, M., Yasir, M., Afzal, S., Aadil, F., Selim, M. M., & Muhammad, K. (2020). A local and global event sentiment based efficient stock exchange forecasting using deep learning. International Journal of Information Management, 50, 432–451. https://doi.org/10.1016/j.ijinfomgt.2019.07.011
Nelson, D. M. Q., Pereira, A. C. M., & de Oliveira, R. A. (2017). Stock market’s price movement prediction with LSTM neural networks. 2017 International Joint Conference on Neural Networks (IJCNN), 1419–1426. https://doi.org/10.1109/IJCNN.2017.7966019
Qaiser, S., & Ali, R. (2018). Text Mining: Use of TF-IDF to Examine the Relevance of Words to Documents. International Journal of Computer Applications, 181. https://doi.org/10.5120/ijca2018917395
Ranjit, S., Shrestha, S., Subedi, S., & Shakya, S. (2018). Comparison of algorithms in Foreign Exchange Rate Prediction. 2018 IEEE 3rd International Conference on Computing, Communication and Security (ICCCS), 9–13. https://doi.org/10.1109/CCCS.2018.8586826
Rosa, C. (2013). The Financial Market Effect of FOMC Minutes (SSRN Scholarly Paper ID 2378398). Social Science Research Network. https://papers.ssrn.com/abstract=2378398
Tadle, R. C. (2021). FOMC minutes sentiments and their impact on financial markets. Journal of Economics and Business, 106021. https://doi.org/10.1016/j.jeconbus.2021.106021
Tf-idf. (2020). 收入維基百科，自由的百科全書. https://zh.wikipedia.org/w/index.php?title=Tf-idf&oldid=63149410

指導教授

陳以錚(Yi-Cheng Chen)

審核日期

2021-9-2

推文