Polite Personal Agents.

The article reports on the development of a user-profiling strategy for intelligent agents that personalizes and improves the interaction between a user and his or her personal agent. Contextual data were captured and modeled to enable the agent to learn how to behave and interact in various situations. The strategy was contrasted with the usual confidence-based method in the calendar management domain. On the average, the proposed strategy had a higher overall percentage of correct assistance actions, assistance interactions, or interactions compared to the confidence-based algorithm.