S. Bontorin,S. Centellegher,R. Gallotti,L. Pappalardo,B. Lepri,& M. Luca, Mixing individual and collective behaviors to predict out-of-routine mobility, Proc. Natl. Acad. Sci
Objective:
- Introduce an approach for next-location prediction that dynamically integrates individual and collective mobility behaviours, leveraging the expected predictability of an individual’s movement
Case:
- 5 Cities in the US
Methodology:
- Deep learning
- Markov chain
Data Source
- GPS location: Filter users with fewer than two trajectories and remove the top 95th percentile of the most represented users
Findings:
