- 學位論文
利用生成對抗綱路預測股票價格
Stock Price Prediction Using Generative Adversarial Networks
若您是本文的作者，可授權文章由華藝線上圖書館中協助推廣。
摘要
近年來深度學習以強大的數據處理能力被廣泛地運用在金融市場中，如股市預測、高頻交易、投資組合優化和交易策略等。其中具有時間序列資料的股票市場因為高雜訊、非穩態、非線性的特徵成為深度學習熱門的研究領域之一。 本論文使用生成對抗網路架構GAN和WGAN進行股票預測，其中使用雙向長短記憶網路(Bi-LSTM)作為生成模型，用以輸入歷史資料以生成未來一天的股票資料，並使用卷積神經網路(CNN)作為判別器，用以判斷生成股票資料與真實股票資料之間的相似性。 本論文以台股作為標的使用生成對抗網路訓練Bi-LSTM模型，研究生成對抗網路在處理數據方面的能力，再以GAN和WGAN比較LSTM模型，研究結果顯示使用生成對抗網路架構的模型可以提升股價的預測表現，但GAN和WGAN在預測股價的表現不分軒輊。
並列摘要
In recent years, deep learning has been widely used in financial market due to its powerful data analysis capabilities. It has been used in various applications such as stock market prediction, high-frequency trading, portfolio optimization, and trading strategies. Stock price, which involves time series data and exhibits high noise, non-stationarity, and non-linearity, has become a popular research area for deep learning. In this thesis, it proposes Generative Adversarial Network (GAN) architecture for stock prediction. Using Bi-LSTM as the generator model to input historical data and generate future stock data for the next day. Convolutional Neural Network (CNN) is utilized as the discriminator to distinguish the generated stock data and real stock data. The thesis focuses on the Taiwan stock market and trains the Bi-LSTM model using the GAN architecture. It discusses the data analysis capabilities of the generative network. Furthermore, it compares GAN and Wasserstein GAN (WGAN) with the LSTM model in terms of predicting stock prices. Results showed that the model using GAN architecture can improve the performance of stock price prediction. However, there is no significant difference between GAN and WGAN in terms of predicting stock prices.
參考文獻
[1] Lo, A. W., & MacKinlay, A. C. (1988). Stock market prices do not follow random walks: Evidence from a simple specification test. Review of Financial Studies, 1(1), 41–66.
[2] Fama, E. F. (1965). The behavior of stock-market prices. Journal of Business, 38(1), 34–105.
[3] Malkiel, B. G. (2003). The efficient market hypothesis and its critics. Journal of Economic Perspectives, 17(1), 59–82.
[4] Patel, J., Shah, S., Thakkar, P., & Kotecha, K. (2015). Predicting stock and stock price index movement using trend deterministic data preparation and machine learning techniques. Expert Systems with Applications, 42(1), 259–268.
[5] LeCun, Y., Boser, B., Denker, J. S., Henderson, D., Howard, R. E., Hubbard, W., & Jackel, L. D. (1989). Backpropagation applied to handwritten zip code recognition. Neural Computation, 1(4), 541–551.