GTIN: Unified Framework for Joint Event and Time Prediction in Temporal Graphs
A new mathematical framework called GTIN has been developed by researchers to simultaneously forecast the next event and its timing within temporal graphs. These graphs represent dynamic systems found in areas like social networks, financial networks, and traffic networks. GTIN is crafted to handle diverse complexities present in temporal graphs, making it adaptable to various network structures and temporal behaviors. Evaluations conducted on several datasets reveal that GTIN surpasses current methods, especially in cases with irregular event patterns and intricate temporal relationships. The research paper can be accessed on arXiv under the identifier 2607.23556.
Key facts
- GTIN is a unified framework for joint event and time prediction in temporal graphs.
- Temporal graphs model dynamic systems in social, financial, and traffic networks.
- The framework captures varying degrees of complexity across temporal graphs.
- It accommodates a wide range of network structures and temporal dynamics.
- Empirical evaluations show GTIN outperforms existing techniques.
- GTIN is particularly effective for irregular event patterns and complex temporal dependencies.
- The paper is available on arXiv with identifier 2607.23556.
- The announcement type is replace-cross.
Entities
Institutions
- arXiv