Abstract:
Human-Computer collaborations abound with uncertainties of various kinds.
Apart from communication languages, the collaborations rely on the
collaborative experience between human and computer, or human and human.
A user behavior model is an instance of experience at a computer, and the
process of building this model is a mutual collaborative process. In this paper,
an assistant agent is described to improve the human-computer collaboration
in an e-Market system. This agent analyzes the system historical data as a
source of its experience, and builds a user behavior model. It uses a modified
association rules mining to discover interrelationships among frequencies and
time sequences of various behaviors. As a result of understandings between
computers and humans, plans associated with goals are recognized.
