Impersonation fraud detection on building access control systems: An approach based on anomalous social and spatio-temporal behaviors

Gabriel Mariano de Castro SIlva, Jaime Simão Sichman

Abstract: Highlights•Goal: pattern of user mobility to eventually detect outlier trajectories.•Motivation: outliers could possibly indicate impersonation fraud.•Focus: access control systems, where checkpoints gather users’ spatio-temporal data.•Novelty: use of group trajectories to enhance the user profile.•Result: detector performance enhanced in a real smart building experiment.