About this document
Real-Time Traffic Accident Prediction Model by Eduardo Moreira is a document available to read on EtoBox.
The document presents SSL-eKamba, a self-supervised framework for real-time traffic accident prediction that addresses challenges in generalization and computational efficiency. It utilizes two auxiliary tasks to enhance traffic pattern representation and redesigns the Kolmogorov-Arnold Network architecture for improved performance. Extensive experiments demonstrate that SSL-eKamba outperforms existing models, offering insights for other spatiotemporal tasks.
- Author
- Eduardo Moreira
- Language
- EN