AI-based stock market trending analysis

• Main Author: Tan, Jess Jing Yi
• Other Authors: Li Fang
• Format: Final Year Project
• Language: English
• Published: Nanyang Technological University 2021
• Subjects: Engineering::Computer science and engineering::Computing methodologies::Artificial intelligence
• Online Access: https://hdl.handle.net/10356/148143
• Institution: Nanyang Technological University

Stock market prediction is widely sought after as the successful prediction could yield rewards of significant profits. With a multitude of factors that affects the value of a stock, the stock market is highly dynamic and seemingly random. The advancement of Artificial Intelligent technology has enabled us to analyse and predict the stock market more effectively and efficiently, as machines are capable to performing calculations beyond human limitations of memory and attention span. The trend of a stock’s price is dependent on the public’s perspective (sentiments) towards it, suggesting the inclusion of sentiment data from sources that could present the public sentiment. This project focuses on improving the prediction performance of the existing work that uses LSTM models to perform the task of stock market prediction, by using a Transformer architecture with the understanding of the concept of time and with the additional feature of news sentiments to enhance the prediction qualities of the model. The proposed methodologies can also generalise to other stocks, suggesting applications beyond the initial scope of this project.