Attention-based security price prediction with transformer fuzzy deep neural network (TFDNN)

This research paper explores the application of Transformer architecture as a Fuzzy Neural Network (FNN) for predicting security prices. The Transformer model, known for its prowess in handling sequence-to-sequence problems, is adapted to accommodate the intricacies of financial data forecasting. Le...

Full description

Saved in:
Bibliographic Details
Main Author: Wijaya, Timothy Larry
Other Authors: Quek Hiok Chai
Format: Final Year Project
Language:English
Published: Nanyang Technological University 2024
Subjects:
Online Access:https://hdl.handle.net/10356/175190
Tags: Add Tag
No Tags, Be the first to tag this record!
Institution: Nanyang Technological University
Language: English
Description
Summary:This research paper explores the application of Transformer architecture as a Fuzzy Neural Network (FNN) for predicting security prices. The Transformer model, known for its prowess in handling sequence-to-sequence problems, is adapted to accommodate the intricacies of financial data forecasting. Leveraging its inherent ability to capture long-range dependencies within sequential data, the Transformer is fine-tuned to operate effectively within the domain of security price prediction. The integration of fuzzy logic within the neural network framework enhances the model’s capability to handle uncertainties inherent in financial markets. Through empirical evaluation on real-world security datasets, the proposed Transformer-based FNN demonstrates promising predictive performance, showcasing its potential as a viable tool for financial forecasting tasks. This study contributes to the ongoing discourse on the fusion of advanced machine learning techniques with financial analysis, offering insights into the utility of Transformer models in addressing complex time-series prediction challenges in the domain of security pricing.