TV program selection system based on a user model agent

摘要

With the coming of the digital broadcasting era, there is now a demand for a "personal information-analysis agent" that can analyze program information collected by another agent according to user needs, and feed the results of analysis back to the user in an easy-to-understand format. In this paper, we describe a user-model reasoning technique that employs a database of viewer stereotypes. This is one type of program-selection function for use in an agent-based TV-program receiving system. To conduct optimal reasoning, the user-model reasoning technique establishes an agent for each stereotype, and predicts what programs would be desirable for a user on the basis of distributed and cooperative processing performed by these stereotype agents.