In developing ubiquitous or pervasive systems it isrnessential that the complexity of the underlying system isrnhidden from the user. To achieve this, the system needs torntake many decisions on behalf of the user. This can onlyrnbe done if the system knows what the user would prefer,rnI.e. It maintains a set of user preferences for each user.rnThis is a laborious task for the user to perform manuallyrnand research is focussing on the use of machine learningrnto assist the user in creating and maintaining anrnacceptable set of preferences. This paper describes howrnstereotypes can be adapted for use in pervasive systems tornhelp build up user preferences while maintaining userrnprivacy through the use of virtual identities, and howrnthese can be modified to match the changing preferencesrnof the group of users who select this stereotype. The paperrnalso introduces the notion of group identities and showsrnhow the same approach can be used to handle these in thernDaidalos pervasive system.
展开▼
