Deep learning methods for financial time series classification

• Main Author: Chua, Dian Lun
• Other Authors: Ponnuthurai Nagaratnam Suganthan
• Format: Final Year Project
• Language: English
• Published: Nanyang Technological University 2021
• Subjects: Engineering::Electrical and electronic engineering
• Online Access: https://hdl.handle.net/10356/149084
• Institution: Nanyang Technological University

With the rapid development in Artificial Intelligence and the rise in financial literacy among people, many are trying to use artificial intelligence as a means to predict the trend of the stock market to increase their wealth more consistently. However, due to the volatility of the stock market, predicting the trend of the stock market remains a daunting task. Therefore, in this paper, we are going to dwell deeper into Artificial Intelligence, specifically into Deep Learning, and use various Deep Learning Models to predict the trend of the stock market as well as discussing the accuracy of the different models. We will be focusing mainly on various Deep Learning models such as Feed Forward Neural Network (FNN), Recurrent Neural Network (RNN), Long Short-Term Memory (LSTM), Gated Recurrent Units (GRUs), Convolutional Neural Network (CNN), Temporal Convolutional Network (TCNs) and Ensembled Deep Random Vector Functional Link (ed-RVFL). In addition, to explore the effects of ensemble learners, we decided to create an ensemble of LSTM, GRU, and TCN. This project aims to use Deep Learning models for Financial Times Series Classification. Based on the experiment with 10 stocks, we found TCN, ed-RVFL, and LSTM-GRU-TCN Ensemble produced better results as compared to the rest of the remaining models.