Details zur Publikation |
| Kategorie | Textpublikation |
| Referenztyp | Tagungsbeiträge |
| URL | https://ceur-ws.org/Vol-3659/ |
Lizenz  |
|
| Titel (primär) | A snapshot-based knowledge graph model for temporal link prediction |
| Titel (sekundär) | Proceedings of the Workshop, Poster and Demonstration Sessions at IJCKG 2023, co-located with 12th International Joint Conference on Knowledge Graphs (IJCKG 2023), Tokyo, Japan, December 8-9, 2023 |
| Autor | Plamper, P.; Lechtenfeld, O.J.
; von Tümpling, W.
; Groß, A. |
| Herausgeber | Yamaguchi, A.; Egami, S.; Kozaki, K.; Kawamura, T.; Villazón-Terrazas, B.; Buranarach, M. |
| Quelle | CEUR Workshop Proceedings |
| Erscheinungsjahr | 2024 |
| Department | FLOEK; EAC |
| Band/Volume | 3659 |
| Seite von | 64 |
| Seite bis | 79 |
| Sprache | englisch |
| Topic | T9 Healthy Planet T5 Future Landscapes T4 Coastal System |
| Abstract | Many systems can be intuitively modeled as knowledge graphs using entities and their relationships. However, we often have only partial or little knowledge of the inherent processes of complex, changing systems such as biomedical, economic or ecological systems. As a result, the construction of knowledge graphs often suffers from incompleteness which can lead to inaccurate analysis results and incorrect conclusions. A widely used approach is to monitor and analyse complex changing systems using time series of measurements. To understand a complex temporal network of processes, it is crucial to identify inherent temporal relationships and interactions. A complete temporal knowledge graph model could provide a better foundation for applications in complex systems and increase its potential to add context and connections that allow uncovering hidden or unknown relationships in data. We propose a snapshot-based knowledge graph model and temporal link prediction algorithm to find relationships between examined objects in successive time points of multivariate time series. We evaluate and demonstrate the functionality in an environmental chemistry use case and predict the transformations of molecules for two datasets. Our approach is able to discover previously unknown relationships in a snapshot-based knowledge graph helping to better understand the dynamics of the examined system. |
| dauerhafte UFZ-Verlinkung | https://www.ufz.de/index.php?en=20939&ufzPublicationIdentifier=30300 |
| Plamper, P., Lechtenfeld, O.J., von Tümpling, W., Groß, A. (2024): A snapshot-based knowledge graph model for temporal link prediction In: Yamaguchi, A., Egami, S., Kozaki, K., Kawamura, T., Villazón-Terrazas, B., Buranarach, M. (eds.) Proceedings of the Workshop, Poster and Demonstration Sessions at IJCKG 2023, co-located with 12th International Joint Conference on Knowledge Graphs (IJCKG 2023), Tokyo, Japan, December 8-9, 2023 CEUR Workshop Proceedings 3659 Rheinisch-Westfälische Technische Hochschule (RWTH), Aachen, 64 - 79 |
|