The Role of Support and Confidence in Collaborative Query Answering
Zbigniew W. Raś · Advances in intelligent and soft computing · 2001
In this paper we show how support and confidence of agent requests for definitions of attribute values (rules) can be used by a distributed society of collaborating agents to improve query answering. Query answering system linked with each agent is different from solving queries by a conventional distributed database or cooperative database in the sense that it discovers rules, if needed, and uses them to resolve foreign attributes in a query. Each agent is seen either as a knowledge system coupled with a query-answering engine (see [8,9,101) or an extended relational DB (usually incomplete) coupled with a query answering engine. The quality of a query answering system of an agent is dependent on definitions (knowledge) he receives from other agents. If definitions of foreign attributes to an agent are more precise, then the new query replacing the original query may give more precise answer. Both support and confidence of agents requests for definitions (rules) can be computed and used to decide which agents should be contacted first to answer queries or if needed asked for additional help.