|Deep learning for financial market prediction
|Chi, Zheru (EIE)
|Stock price forecasting
Hong Kong Polytechnic University -- Dissertations
|Department of Electronic and Information Engineering
|viii, 64 pages : color illustrations
|In order to effectively predict the trend of stock prices, a method is proposed by using fundamental analysis and deep learning algorithm. Firstly, S&P 500 data from 1980 to 2019 is collected and analyzed via Yahoo Finance API and the logarithmic return rate, volatility and trading volume on different dates are obtained. Secondly, the S&P 500 from 1980 to 2019 is also converted into image data. The quantitative data is converted into f-line pictures, and the daily trading volume histogram is merged to obtain 224×224×3 images. Thirdly, several deep learning algorithms, such as Long Short-Term Memory (LSTM) and convolutional neural network (CNN), are compared to determine the trend of stock price after 21 days through parameter optimization. The classification results show that the CNN model with quantitative data as input has strong robustness and obtains the classification results (66.70% in accuracy and 0.5744 in AUC). Finally, by randomly removing some tags and improving the training model, the maximum accuracy of 78.37% and the best AUC of 0.73 can be obtained.
|All rights reserved
Files in This Item:
|For All Users (off-campus access for PolyU Staff & Students only)
As a bona fide Library user, I declare that:
- I will abide by the rules and legal ordinances governing copyright regarding the use of the Database.
- I will use the Database for the purpose of my research or private study only and not for circulation or further reproduction or any other purpose.
- I agree to indemnify and hold the University harmless from and against any loss, damage, cost, liability or expenses arising from copyright infringement or unauthorized usage.
By downloading any item(s) listed above, you acknowledge that you have read and understood the copyright undertaking as stated above, and agree to be bound by all of its terms.
Please use this identifier to cite or link to this item: