Tag
This paper proposes the Activity Chain Encoder (ACE), a self-supervised model that learns daily activity patterns from mobile phone location data without labels, and demonstrates its effectiveness in identifying differentiated activity patterns in urban environments like London.