Responsive image
博碩士論文 etd-0716117-145700 詳細資訊
Title page for etd-0716117-145700
論文名稱
Title
使用深度學習卷積神經網路預測股票買賣策略之分類研究
Applying Convolution Neural Network in Deep Learning to Predict on Stock Trading Strategy
系所名稱
Department
畢業學年期
Year, semester
語文別
Language
學位類別
Degree
頁數
Number of pages
74
研究生
Author
指導教授
Advisor
召集委員
Convenor
口試委員
Advisory Committee
口試日期
Date of Exam
2017-07-31
繳交日期
Date of Submission
2017-08-16
關鍵字
Keywords
卷積神經網路、交易策略、人工智慧
Convolutional Neural Network, Artificial Intelligence, Stock investment
統計
Statistics
本論文已被瀏覽 6041 次,被下載 744
The thesis/dissertation has been browsed 6041 times, has been downloaded 744 times.
中文摘要
電腦硬體的快速發展,擁有驚人的處理量及速度,讓需要經過大量運算的人工智慧再次崛起,而深度學習是人工智慧中成長最為快速的領域,本研究使用深度學習的卷積神經網路當作架構,針對台灣50指數ETF預測,建立臺灣50指數ETF的買賣時機建議模型,透過卷積神經網路自我學習特徵的方式找出買賣的時機,本研究對於卷積神經網路模型預測股票資料提出參數的優化,(1)卷積層數,(2)訓練資料長短,(3)卷積核大小,(4)買賣策略,進行分析找出最好的參數及組合。

本研究也提出兩階層架構的方式來預測買賣策略,(1)先進行股票多頭空頭市場的預測 (2) 再預測股票買賣策略,實驗結果驗證兩階層架構相比一階層架構在某些情況下會有較好的效果。

為了驗證預測買賣的效果,除了準確率之外,本研究採用投資報酬率和相同交易期間的傳統投資策略「買入及持有(buy and hold)」比較,結果顯示本系統的研究模型超越相同期間的傳統投資策略,說明卷積神經網路在股票預測上有不錯的效果。
Abstract
The rapid increase of computing power, along with the improvement of software capabilities has made artificial intelligence a new trend. Deep learning recently becomes the fastest growing area in artificial intelligence. This study examines how the deep convolution neural network (CNN) can be applied to predict stock investment and evaluated with the Taiwan 50 Index ETF data. Four parameters related to the performance of CNN in stock investment are investigated in our experiment: (1) Number of convolution layers , (2) Length of training period ,(3) Convolution size , (4) Trading strategy.

The study also proposes a two-tier model for stock trading, (1) the first predicts the overall market trend (bull or bear) and (2) the second predicts the stock trading strategy (buy, hold or sell). The performance of different strategies is evaluated on testing datasets of 2015 and 2016. The return of buy-and-hold strategy is used as the benchmark for comparison. Our findings indicate that CNN can perform significantly better than the benchmark, but the exact results vary under different learning parameters. The result also shows that the two-tier architecture performed better in some cases.
目次 Table of Contents
論文審定書 .......................................................... i
致謝 ............................................................... ii
摘要 .............................................................. iii
Abstract ........................................................... iv
圖目錄 ............................................................ vii
表目錄 ........................................................... viii
第一章 緒論 ...................................................... 1
第一節 研究背景 .......................................................... 1
第二節 研究動機與目的 .................................................... 2
第三節 研究流程 .......................................................... 3
第二章 文獻探討 .................................................. 5
第一節 類神經網路預測股票相關研究 ...................................... 5
第二節 卷積神經網路相關研究 ............................................ 7
第三節 卷積神經網路 .................................................... 8
第四節 卷積神經網路的模型架構 ......................................... 10
第五節 Resnet架構 ..................................................... 14
第三章 研究方法 ................................................. 17
第一節 導論 ........................................................... 17
第二節 研究資料與實驗環境 ............................................. 18
第三節 卷積神經網路參數影響 ........................................... 19
第四節 變數的挑選 ..................................................... 19
第五節 變數轉成特徵矩陣 ............................................... 20
第六節 多頭空頭市場的預測 ............................................. 25
第七節 預測投資策略 ................................................... 25
第四章 研究結果及分析 ........................................... 27
第一節 實驗驗證及說明 ................................................. 27
第二節 卷積模型因素 ................................................... 30
第三節 買賣策略因素 ................................................... 38
第四節 資料因素 ....................................................... 40
第五節 多頭空頭市場預測 ............................................... 53
第五章 結論與建議 ............................................... 57
第一節 研究結論 ....................................................... 57
第二節 研究限制與建議 ................................................. 60
參考文獻 ........................................................... 61
中文部分 ................................................................ 61
英文部分 ................................................................ 62
參考文獻 References
中文部分:
1.尤韻涵. (2009). 台股指數開盤價格之預測-應用類神經網路及灰預測模型. (碩士), 輔仁大學, 新北市.
2.洪文麟. (2016). 深度學習應用於以影像辨識為基礎的個人化推薦系統-以服飾樣式為例. (碩士), 國立成功大學, 台南市.
3.曹瑋宸. (2016). 基於深度學習之性別辨識與人員計數. (碩士), 國立雲林科技大學, 雲林縣.
4.陳家隆. (2002). 運用統計方法與人工智慧技術建構整合性投資策略. (碩士), 國立成功大學, 台南市.
5.黃信凱. (2016). 深度摺積神經網路於混合式整體學習之影像檢索技術. (碩士), 國立中央大學, 桃園縣.
6.蔣岡霖. (2000). 應用股價技術圖型比對分析預測未來股價趨勢. (碩士), 大葉大學, 彰化縣.
7.盧昭亮. (2016). 基於卷積神經網路之息肉狀脈絡膜血管病變偵測. (碩士), 國立中正大學, 嘉義縣.

英文部分
1. Bekiros, S. D., & Georgoutsos, D. A. (2008). Direction‐of‐change forecasting using a volatility‐based recurrent neural network. Journal of Forecasting, 27(5), 407-417.
2. Cao, Q., Parry, M. E., & Leggio, K. B. (2011). The three-factor model and artificial neural networks: predicting stock price movement in China. Annals of Operations Research, 185(1), 25-44.
3. De Faria, E., Albuquerque, M. P., Gonzalez, J., Cavalcante, J., & Albuquerque, M. P. (2009). Predicting the Brazilian stock market through neural networks and adaptive exponential smoothing methods. Expert Systems with Applications, 36(10), 12506-12509.
4. Ding, X., Zhang, Y., Liu, T., & Duan, J. (2015). Deep Learning for Event-Driven Stock Prediction. Paper presented at the Ijcai.
5. Fabozzi, F. J., & Francis, J. C. (1977). Stability tests for alphas and betas over bull and bear market conditions. The Journal of Finance, 32(4), 1093-1099.
6. Güresen, E., & Kayakutlu, G. (2008). Forecasting Stock Exchange Movements Using Artificial Neural Network Models and Hybrid Models. Intelligent Information Processing, 288, 129-137.
7. He, K., Zhang, X., Ren, S., & Sun, J. (2016). Deep residual learning for image recognition. Paper presented at the Proceedings of the IEEE conference on computer vision and pattern recognition.
8. Hubel, D. H., & Wiesel, T. N. (1962). Receptive fields, binocular interaction and functional architecture in the cat's visual cortex. The Journal of physiology, 160(1), 106-154.
9. Hung, S.-Y., Liang, T.-P., & Liu, V. W.-C. (1996). Integrating arbitrage pricing theory and artificial neural networks to support portfolio management. Decision Support Systems, 18(3-4), 301-316.
10. Krizhevsky, A., Sutskever, I., & Hinton, G. E. (2012). Imagenet classification with deep convolutional neural networks. Paper presented at the Advances in neural information processing systems.
11. LeCun, Y., Bottou, L., Bengio, Y., & Haffner, P. (1998). Gradient-based learning applied to document recognition. Proceedings of the IEEE, 86(11), 2278-2324.
12. Simonyan, K., & Zisserman, A. (2014). Very deep convolutional networks for large-scale image recognition. arXiv preprint arXiv:1409.1556.
13. Srivastava, R. K., Greff, K., & Schmidhuber, J. (2015). Highway networks. arXiv preprint arXiv:1505.00387.
14. Szegedy, C., Liu, W., Jia, Y., Sermanet, P., Reed, S., Anguelov, D., . . . Rabinovich, A. (2015). Going deeper with convolutions. Paper presented at the Proceedings of the IEEE conference on computer vision and pattern recognition.
15. Tilakaratne, C., Mammadov, M. A., & Morris, S. A. (2009). Modified neural network algorithms for predicting trading signals of stock market indices. Advances in Decision Sciences, 2009.
電子全文 Fulltext
本電子全文僅授權使用者為學術研究之目的,進行個人非營利性質之檢索、閱讀、列印。請遵守中華民國著作權法之相關規定,切勿任意重製、散佈、改作、轉貼、播送,以免觸法。
論文使用權限 Thesis access permission:自定論文開放時間 user define
開放時間 Available:
校內 Campus: 已公開 available
校外 Off-campus: 已公開 available


紙本論文 Printed copies
紙本論文的公開資訊在102學年度以後相對較為完整。如果需要查詢101學年度以前的紙本論文公開資訊,請聯繫圖資處紙本論文服務櫃台。如有不便之處敬請見諒。
開放時間 available 已公開 available

QR Code