跳到主要內容

簡易檢索 / 詳目顯示

研究生: 陳思奇
Chen, Si-Qi
論文名稱: 使用C-RNN神經網絡模型預測匯率變動—以中美日台為例
Using C-RNN Neural Network Model to Predict Exchange Rate Movements - A Case Study of China, America, Japan and Taiwan
指導教授: 廖四郎
Liao, Szu-Lang
口試委員: 林建秀
Lin, Chien-Hsiu
連育民
Lan, Yu-Min
黃星華
Huang, Hsing-Hua
學位類別: 碩士
Master
系所名稱: 商學院 - 金融學系
Department of Money and Banking
論文出版年: 2020
畢業學年度: 108
語文別: 中文
論文頁數: 29
中文關鍵詞: 深度學習卷積神經網絡循環神經網絡C-RNN匯率
外文關鍵詞: Deep learning, Convolutional neural network, Circular neural network, C-RNN, Exchange rate
DOI URL: http://doi.org/10.6814/NCCU202001114
相關次數: 點閱:129下載:40
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
  • 本篇論文採用了將卷積神經網絡和循環神經網絡相結合的C-RNN模型來作為預測未來匯率價格的工具,希望藉由此工具能預判未來匯率的走勢與價格來作為參考。為此本研究選用了CNY/USD、CNY/ TWD、CNY/JPY等四種貨幣間的三種匯率價格作為分析資料,將未來5天的匯率作為預測目標。C-RNN是一種深度學習的模型,由於其將(CNN)卷積神經網絡和(RNN)循環神經網絡相結合,擁有著兩者的各自優勢,既能從資料中提取出空間特徵又能通過循環掌握時間特徵,因此可能在對匯率的預測上能取得良好成果。


    This paper uses a C-RNN model that combines convolutional neural networks and recurrent neural networks as a tool to predict future exchange rate. It is hoped that this tool can predict future exchange rate trends and prices as a reference. For this reason, this study selected three exchange rates among four currencies such as CNY/USD, CNY/TWD, and CNY/JPY as analysis data, and the exchange rate for the next 5 days was used as the forecast target. C-RNN is a deep learning model. Because it combines (CNN) Convolutional Neural Network and (RNN) Recurrent Neural Network, it has their own advantages. It can extract spatial features and time characteristics from data at the same time, so it is possible to achieve good results in the forecast of exchange rates.

    第一章 緒論 1
    第一節 研究動機 1
    第二節 研究目的 2
    第三節 研究背景 2
    第二章 文獻探討 5
    第一節 神經網絡模型應用之相關文獻 5
    第二節 文獻回顧總結 5
    第三章 研究方法 6
    第一節 卷積神經網絡 6
    第二節 循環神經網絡 10
    第三節 卷積-循環神經網絡 13
    第四節 研究對象 14
    第五節 一維時間序列資料二維化 14
    第四章 實證分析 16
    第一節 實驗架構設計 16
    第二節 實證結果 21
    第五章 結論與建議 27
    第一節 結論 27
    第二節 未來展望 27
    參考文獻 28

    [1] Chen, K., Zhou, Y., & Dai, F. (2015, October). A LSTM-based method for stock returns prediction: A case study of China stock market. In 2015 IEEE international conference on big data (big data) (pp. 2823-2824). IEEE.
    [2] Dunis, C. L., & Huang, X. (2002). Forecasting and trading currency volatility: An application of recurrent neural regression and model combination. Journal of forecasting, 21(5), 317-354.
    [3] Dunis, C. L., Laws, J., & Sermpinis, G. (2011). Higher order and recurrent neural architectures for trading the EUR/USD exchange rate. Quantitative Finance, 11(4), 615-629.
    [4] Guresen, E., Kayakutlu, G., & Daim, T. U. (2011). Using artificial neural network models in stock market index prediction. Expert Systems with Applications, 38(8), 10389-10397.
    [5] Gao, S. E., Lin, B. S., & Wang, C. M. (2018, December). Share price trend prediction using CRNN with LSTM structure. In 2018 International Symposium on Computer, Consumer and Control (IS3C) (pp. 10-13). IEEE.
    [6] Takeuchi, L., & Lee, Y. Y. A. (2013). Applying deep learning to enhance momentum trading strategies in stocks. In Technical Report. Stanford University.
    [7] Tino, P., Schittenkopf, C., & Dorffner, G. (2001). Financial volatility trading using recurrent neural networks. IEEE Transactions on Neural Networks, 12(4), 865-874.
    [8] Yu, S. S., Chu, S. W., Chan, Y. K., & Wang, C. M. (2019). Share Price Trend Prediction Using CRNN with LSTM Structure. Smart Science, 7(3), 189-197.
    [9] 賴嘉蔚,(2018)。卷積神經網絡預測時間序列能力分析。國立政治大學金融學研究所碩士論文,台北市。取自https://hdl.handle.net/11296/y25ux2

    QR CODE
    :::