Title
Modelling of Behavioural Patterns for Abnormality Detection in the Context of Lifestyle Reassurance
Abstract
As a consequence of the growing number of older and vulnerable people, health and care providers are increasingly considering new approaches to support people in their own homes. In this context, lifestyle reassurance analyses data collected from a range of sensors to determine a person's `routine' and highlights any important changes. This paper proposes a new approach for detection of individual deviation from normal behaviour focusing on building probabilistic models of behaviour based on a set of activity attributes. Models are trained using only normal behaviour. Variations from the models are considered as abnormal behaviours and these can be highlighted for subsequent review or intervention. Case study experiments with real life data suggest that some users' activities follow regular patterns and that these patterns can be learned with probabilistic models.
Year
DOI
Venue
2008
10.1007/978-3-540-85920-8_30
CIARP
Keywords
Field
DocType
lifestyle reassurance analyses data,new approach,activity attribute,lifestyle reassurance,probabilistic model,behavioural patterns,vulnerable people,normal behaviour,real life data,case study experiment,abnormal behaviour,care provider,abnormality detection,data collection
Pattern recognition,Computer science,Cognitive psychology,Telecare,Normal behaviour,Artificial intelligence,Abnormality detection,Probabilistic logic
Conference
Volume
ISSN
Citations 
5197
0302-9743
12
PageRank 
References 
Authors
0.72
10
4
Name
Order
Citations
PageRank
Fabien Cardinaux127919.00
Simon Brownsell2283.85
Mark Hawley3141.15
D A Bradley4466.39