Early Warning Detection For Extreme Events In Time Series Using Topological Data Analysis
| dc.contributor.author | KHOO ZI YIK | |
| dc.date.accessioned | 2026-04-09T07:44:20Z | |
| dc.date.issued | 2025 | |
| dc.description | This research investigates the application of an approach in Topological Data Analysis (TDA), specifically persistent homology, to detect extreme events in time series data as an early warning signal. Extreme events, often occurring unexpectedly and without prior warning, can have significant repercussions on economic and social systems. These events are frequently derived from time series data, which pose analytical challenges due to their inherent characteristics such as non-linearity, noise, and fluctuations. These complexities complicate the use of traditional statistical methods for effective analysis. The most recent and impactful extreme event in Malaysia was the COVID-19 pandemic. Consequently, this study focuses on COVID-19 as a case study for detecting extreme events. The research outlines a systematic methodology, beginning with data collection and pre-processing through phase plane reconstruction, followed by feature extraction using persistent homology, and culminating in feature selection. This methodological framework aims to provide valuable insights into the underlying structure of time series data and to facilitate the identification of anomalies or extreme events within the data. | |
| dc.identifier.uri | https://scholarhub.unimas.my/handle/123456789/346 | |
| dc.language.iso | English | |
| dc.publisher | Universiti Malaysia Sarawak (UNIMAS) | |
| dc.relation.ispartofseries | Faculty of Computer Science and Information Technology | |
| dc.subject | Early warning, extreme events, persistent homology, time series data, topological data analysis (TDA) | |
| dc.title | Early Warning Detection For Extreme Events In Time Series Using Topological Data Analysis | |
| dc.type | Final Year Project |