研究生: |
鄭邦廷 Cheng, Pan-Ting |
---|---|
論文名稱: |
基於深度學習與技術分析指標預測股市買賣點 Stock buy and sell points prediction based on deep learning and technical analysis indicators |
指導教授: |
吳順德
Wu, Shuen-De |
口試委員: |
呂有勝
Lu, Yu-Sheng 劉益宏 Liu, Yi-Hung 吳順德 Wu, Shuen-De |
口試日期: | 2023/07/13 |
學位類別: |
碩士 Master |
系所名稱: |
機電工程學系 Department of Mechatronic Engineering |
論文出版年: | 2023 |
畢業學年度: | 111 |
語文別: | 中文 |
論文頁數: | 66 |
中文關鍵詞: | 機器學習 、人工智慧 、技術分析 |
英文關鍵詞: | Machine Learning, Artificial Intelligence, Technical Analysis |
研究方法: | 比較研究 |
DOI URL: | http://doi.org/10.6345/NTNU202300855 |
論文種類: | 學術論文 |
相關次數: | 點閱:92 下載:10 |
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 |
股票交易市場是由各種金融機構和投資者組成,用於買賣股票和其他金融產品的交易活動。市場的主要目的是提供一個公平、透明和有保障的交易環境,促進股票和其他金融產品的流通。而參與者包括投資者、證券公司、投資銀行、基金、保險公司、政府機構等。投資者可以通過股票交易市場進行交易,包括股票、債券、期貨等各種金融產品。證券公司、投資銀行等機構則負責為投資者提供證券交易相關的服務和產品。
然而投資並不是穩賺不賠,以股票來說,其價值可能伴隨著公司業績表現、總體經濟變化,乃至各種政治因素而有波動。這代表著投資者必須去注意這些市場上的各種外在條件去對自己手中的投資標的去做調整,並無一精準的判斷條件。縱使在許多的外在條件影響之下,歷年仍有許多研究希望能藉由各種方式來判斷進場買進和獲利了結的方法甚至到未來股價的判斷。股票交易的獲利方式在於買進及賣出所產生的價差,但每個投資者所買進及賣出的位置不同,即代表各項因素不同,因此買的價格過高則會有套牢的風險,或者買進後遲遲沒有往上漲,因而賠上交易和時間成本。因此,如果可以藉由調整進出場條件及發動策略去預測未來走向,使我們可以更敏銳的判斷價格變化,就可以在價格即將變化時進場,並且賺取其中價差。
在各種因素影響的情況之下,股票的進出場訊號可視為一非線性的時序訊號。而人工智慧在非線性模型的表現相當優秀,尤其是在處理大量複雜的數據時更顯突出。本論文希望配合著股票的技術分析及回測過的數據結合人工智慧達到預測未來買賣點位之目的,以強化並優化進出條件,進而增進投資報酬率。
The stock market is composed of various financial institutions and investors, used for buying and selling stocks and other financial products. The main purpose of the market is to provide a fair, transparent, and secure trading environment, promote the circulation of stocks and other financial products. Participants include investors, securities firms, investment banks, funds, insurance companies, government agencies, etc. Investors can trade through the stock trading market, including stocks, bonds, futures, and other financial products. Institutions such as securities firms and investment banks are responsible for providing securities trading-related services and products to investors.
However, investing is not always profitable. For example, in the case of stocks, their value may fluctuate along with the company's performance, macroeconomic changes, and even political factors. This means that investors must pay attention to various external factors in the market and adjust their investment targets accordingly, but there is no precise judgment criteria. Despite the many external factors affecting the market, many studies over the years have hoped to find ways to determine when to buy in and take profit even to predict future stock prices. The profit model of stock trading lies in the price difference between buying and selling, but each investor's buying and selling position is different, which means that various factors are different. If the buying price is too high, there is a risk of being locked in, or the entry price may not respond, resulting in transaction and time costs. Therefore, if we can use strategies to adjust entry and exit conditions to predict future trends, we can make more accurate judgments about price changes and enter the market when the price is about to change, and profit from the price difference.
Under the influence of various factors, the entry and exit signals of stocks can be seen as a non-linear signal. Artificial intelligence has shown outstanding performance in non-linear models, especially when dealing with large and complex data. This paper aims to combine technical analysis of stocks with backtested data and artificial intelligence to predict future buy and sell points, and to strengthen and optimize entry and exit conditions.
E. Guresen, G. Kayakutlu and T. U. Daim, “Using artificial neural network models in stock market index prediction,” Expert Systems with Applications, Volume 38, Issue 8, Pages 10389-10397, 2011.
R. Singh and S. Srivastava, “Stock prediction using deep learning,” Multimedia Tools and Applications, Pages 18569-18584, 2017.
A. M. Turing, “I.—COMPUTING MACHINERY AND INTELLIGENCE,” Mind, Volume LIX, Issue 236, Pages 433–460, 1950.
E. Chong, C. Han and F. C. Park, “Deep learning networks for stock market analysis and prediction: Methodology, data representations, and case studies,” Expert Systems with Applications, Volume 83, Pages 187-205, 2017.
H. C. Chia, “Stock market forecasting based on Convolutional Neural Networks and Stacked LSTM combined with shortcut connection,” NTNU, Graduate thesis, 2019.
L. Wenjie, L. Jiazheng, L. Yifan, S. Aijun and W. Jingyang, “A CNN-LSTM-Based Model to Forecast Stock Prices,” Natural Sciences, Pages 1-10, 2020.
I. A. Basheer and M. Hajmeer, “Artificial neural networks: fundamentals, computing, design,gnd application,” Journal of Microbiological Methods, Pages 3-31, 2000.
C. C. Hao, “Low-Cost and Fast Design of Precise Activation Functions in Neural Network,” NCUE, Graduate thesis, 2020.
C. C. Chi, “A relevant loss function in option valuation- Hedging performance,”NCTU, Graduate thesis, 2004.
R. David, H. Geoffrey and W. Ronald, “Learning representations by back-propagating errors,” Nature, Page 533-536, 1986.
A. Shun, “Backpropagation and stochastic gradient descent method,” Neurocomputing 5, Page 185-196, 1993.
R. Janosh, Tikz.net, https://tikz.net/dropout/.
H. J. Yuan, “Development of MLP based Deep Learning Associative Memory Neural Networks and Its Applications,”NFU, Graduate thesis, 2017.
A. Sherstinsky, “Fundamentals of Recurrent Neural Network (RNN) and Long Short-Term Memory (LSTM) network,” Physica D: Nonlinear Phenomena, Volume 404, Page 132306, 2020.
B. Jason, 2017, https://machinelearningmastery.com/gentle-introduction-long-short-term-memory-networks-experts/.
Y. James, 2017, https://medium.com/.
S. Mehtab and J. Sen, “Stock Price Prediction Using Convolutional Neural Networks on a Multivariate Time Series,” Proceedings of the 3rd National Conference on Machine Learning and Artificial Intelligence, Page 7, 2020.
Cmoney, “KD指標是什麼?一分鐘看懂KD線算法!” https://www.cmoney.tw/learn/course/technicals/topic/484.
富途牛牛, https://support.futunn.com/hant/topic658/.
Cmoney, ” MACD 指標意思是什麼?” https://www.cmoney.tw/learn/course/technicals/topic/750.