Behavioural profiling authentication based on trajectory based anomaly detection model of user’s mobility

Abstract

Behavioural profiling and biometry are an interesting concept connected with authentication that have appeared in scientific literature and business world. Those methods indisputably offer new possibilities such as constant authentication and multi-user classification, but their taxonomy and definitions are not as clarified as it is for traditional authentication factors. The approach presented provides in this work provides an example of behavioural authentication model tested on a large dataset, focusing on one aspect of user behaviour - mobility, which can be adjusted to include other aspects in user behavioural authentication model. Also possible applications and extensions to the model are proposed.