About this document
Yeni by Bahadır Altun is a document available to read on EtoBox.
The document presents STEP (STochastic Event Predictor), a novel framework for temporal link prediction that reformulates the problem as a sequential forecasting task rather than binary classification. STEP utilizes discrete temporal motif transitions governed by Poisson processes to model event dynamics and produces compact feature vectors that enhance existing temporal graph neural networks. Experiments demonstrate significant improvements in predictive performance and efficiency over state-of-the-art met
- Author
- Bahadır Altun
- Language
- EN