Date of Award:

12-2024

Document Type:

Thesis

Degree Name:

Master of Science (MS)

Department:

Computer Science

Committee Chair(s)

Shah Muhammad Hamdi

Committee

Shah Muhammad Hamdi

Committee

Soukaina Filali Boubrahimi

Committee

Mahdi Nasrullah Al-Ameen

Abstract

Time series data, such as weather forecasts, stock market trends, or heart rate monitors, plays a vital role in many areas of our lives. However, creating realistic synthetic time series data for research and testing purposes has been a significant challenge due to limitations in existing methods, which often struggle with accuracy and consistency. In this study, we developed two new approaches to generate high-quality time series data more effectively. The first method introduces a dual-feedback system that helps the model learn and replicate real data patterns more accurately by providing guidance at different stages of the learning process. The second method features a novel design that works well with both short and long sequences of data, making it versatile for various applications. Both methods incorporate innovative techniques to improve stability and ensure the generated data closely mirrors real-world patterns. These techniques include algorithms for preserving the best model during training, novel cost functions, and an integrated joint training mechanism to enhance performance. We tested these approaches on different types of time series data, including financial markets and medical signals. The results showed that our methods consistently produced more accurate and reliable data than existing standard techniques.

Share

COinS