USC-InfoLab/busyness-graph-neural-network

Busyness Graph Neural Network (BysGNN): A framework for accurate Point-of-Interest visit forecasting using dynamic graphs that capture spatial, temporal, semantic, and taxonomic contexts. Presented at ACM SIGSPATIAL 2023, this repository includes code, baselines, and experiments.

Language: Jupyter Notebook - Size: 85.5 MB - Last synced at: 6 months ago - Pushed at: 6 months ago - Stars: 1 - Forks: 0