臺灣博碩士論文加值系統

股票和外匯價格的預測是一個具有挑戰性的問題，吸引了研究人員和投資者的廣泛興趣。儘管專家進行了許多研究來應對這一挑戰，但是仍然存在著應對股票和外匯市場價格波動的問題。本文旨在通過使用移動平均線和機器學習技術來減輕外匯市場的波動性對預測的影響。與第二天或下一小時的預測相反，這種方法還可以對整個價格序列產生整體預測。這項研究驗證了這種新穎方法在實際應用中的有效性，從而使最佳性能達到了86.8％的準確度。與其他方法相比，該方法還顯示出最佳的利潤收益。

Stock and Foreign Exchange price prediction is a challenging problem that attracts broad interests from researchers and investors. Despite many studies to tackle this challenge, issues in dealing with the volatility of stock and Forex market prices still remain. This thesis aims to lessen the effect that the volatile nature of the Forex market imposed on predictions by using Moving Averages accompanied by machine learning techniques. This approach also produces an overall prediction for an entire sequence of prices as opposed to next day or next hour prediction. The study verifies the effectiveness of this novel approach on a real-world application, resulting in the best performance reaching 86.8% accuracy. The proposed method also shows the best profit gains compared to other approaches.

Chapter 1. Introduction ------------------ 1
Chapter 2. Related Work ------------------ 3
Chapter 3.Financial Technical Terms ------ 7
3.1 Financial Indicators ----------------- 7
3.1.1 Relative Strength Index ------------ 9
3.1.2 Moving Averages -------------------- 10
3.1.3 Commodity Channel Index ------------ 12
3.1.4 Momentum --------------------------- 13
3.1.5 Mean Deviation --------------------- 14
3.2 Foreign Exchange technical terms ----- 14
Chapter 4. Recurrent Neural Network ----- 17
4.1 LSTM --------------------------------- 19
4.2 GRU ---------------------------------- 22
Chapter 5. Proposed Method --------------- 25
5.1 Future-Past Moving Averages ---------- 26
5.2 Neural Network Architecture ---------- 29
5.3 Performance Index -------------------- 32
Chapter 6. Dataset ----------------------- 35
6.1 Data selection and collection -------- 35
6.2 Data Split --------------------------- 37
Chapter 7. Performance Evaluation -------- 39
7.1 Training ----------------------------- 39
7.2 Accuracy Evaluation ------------------ 39
7.3 Profitability Evaluation ------------- 41
7.4 Comparison --------------------------- 48
Chapter 8. Conclusion/Future works ------- 55
References ------------------------------- 57

[1] Phi, M. (2018) Illustrated Guide to LSTM’s and GRU’s: A step by step explanation
https://towardsdatascience.com/illustrated-guide-to-lstms-and-gru-s-a-step-by-step-explanation-44e9eb85bf21
[2] Bradfield, D. (2019) Forex Vs Stocks: Top Differences and How to Trade Them
https://www.dailyfx.com/education/beginner/forex-vs-stocks.html
[3] Viraf, How (NOT) to Predict Stock Prices with LSTM
https://towardsdatascience.com/how-not-to-predict-stock-prices-with-lstms-a51f564ccbca
[4] Which is Harder to Trade Forex or Stock
https://rockfortmarkets.com/en/which-is-harder-to-trade-forex-or-stocks/
[5] Gregory M., Morris (2006). Candlestick Charting Explained: Timeless Techniques for Trading Stocks and Futures. McGraw-Hill. ISBN 9780071461542.
[6] Nison, Steve (2001). Japanese Candlestick Charting Techniques (2nd ed.). Prentice Hall Press. ISBN 9780735201811.
[7] What are CandleStick Charts
https://www.cmcmarkets.com/en/trading-guides/what-are-candlestick-charts
[8] Kusama, R.M.I., Ho, T., Kao, W., Ou, Y., & Hua,K. (2019) Using Deep Learning Neural Networks and Candlestick Chart Representation to Predict Stock market. arXiv, arXiv-1903. Yuan Ze University, National Taiwan University of Science and Technology
[9] Mudinas, A., Zhang, D., & Levene, M. (2019) Market Trend Prediction using Sentiment Analysis: Lessons Learned and Paths Forward. arXiv preprint arXiv:1903.05440. University of London
[10] Bowling, J. (2019) What Happened When I Tried Market Prediction With Machine Learning
https://towardsdatascience.com/what-happened-when-i-tried-market-prediction-with-machine-learning-4108610b3422
[11] Nison, S, 2001. Japanese candlestick charting techniques, second edition, Upper Saddle River: Prentice Hall Press,.
[12] Weisstein, Eric W. "Mean Deviation." From MathWorld--A Wolfram Web Resource. https://mathworld.wolfram.com/MeanDeviation.html
[13] Britz, D. (2015) Recurrent Neural Networks Tutorial, Part 3 – Backpropagation Through Time and Vanishing Gradients
http://www.wildml.com/2015/10/recurrent-neural-networks-tutorial-part-3-backpropagation-through-time-and-vanishing-gradients/
[14] Hochreiter, S. (1998) The Vanishing Gradient Problem During Learning Recurrent Neural Nets and Problem Solutions. International Journal of Uncertainty Fuzziness and Knowledge-Based Systems , 6(02), 107-116.
[15] J. Chung, C. Gulcehre, K. Cho, Y. Bengio, (2014) Empirical evaluation of gated recurrent neural networks on sequence modeling. arXiv preprint arXiv:1412.3555
[16] Shewalkar, A., Nyavanandi, D. Ludwig, S. A. (2018) Performance Evaluation of Deep Neural Networks Applied To Speech Recognition: RNN, LSTM and GRU. Journal of Artificial Intelligence and Soft Computing Research, 9(4), 235-245. North Dakota State University
[17] Neely, C.J., Weller, P.A. (2011) Technical Analysis in the Foreign Exchange Market. Federal Reserve Bank of St. Louis Working Paper No.
[18] Brouwer, Y., (2017) News Sentiment and Foreign Exchange Markets. Erasmus School of Economics
[19] Roudgar, E. (2012) Forecasting Foreign Exchange Market Trends: Is Technical Analysis Perspective Successful? (Doctoral dissertation, Eastern Mediterranean University (EMU))
[20] G. E. P. Box and G. M. Jenkins, Time Series Analysis: Forecasting and Control, Holden-Day, San Francosco, CA, 2015.
[21] Kamruzzaman, J., and Sarker, R.A., (2004) ANN-Based Forecasting of Foreign Currency Exchange Rates. 3(2), 49-58. Monash University, University of NSW
[22] Rehman, M., Khan, G.M., & Mahmud, S.A. (2014) Foreign Currency Exchange Rates Prediction using CGP and Recurrent Neural Network. IERI Procedia, 10, 239–244. University of Engineering and technology