Résumé
Aging process is associated with behavior change and continuous decline in physical and cognitive abilities. Therefore, early detection of behavior change is major enabler for providing adapted services to elderly people. Today, different psychogeriatric methods target behavior change detection. However, these methods require presence of caregivers and manual analysis. In this paper, we present our technological approach for early behavior change detection. It consists in monitoring and analyzing individual activities using pervasive sensors, as well as detecting possible changes in early stages of their evolution. We also present a first validation of the approach with real data from nursing home deployment.