Cooperative query answering analyzes the intent of a query and provides generalized neighborhood or associated information relevant to the query using the knowledge about the query and data. The main issues of cooperative query answering are knowledge representation at different abstraction levels and supporting inference between those levels. In this paper, we propose the Value and Domain Abstraction Hierarchy (VDAH) as a framework for cooperative query answering. VDAH provides multilevel knowledge representation for both domain abstraction and value abstraction. The maintenance of knowledge about domain abstraction as well as value abstraction can bring out two different advantages - query processing and the management of data and knowledge. We will construct a knowledge base to achieve these advantages.
Based on our model, queries can be classified into four groups - Approximate Selection, Approximate Matching in Selection Operation, Approximate Join, Approximate Matching in Join Operation. In query processing, a relation named ABSTRACTION is used to represent the abstraction relationship between all the values in two different arbitrary domains. Once ABSTRACTION is made, a query can be processed by a natural join. We have demonstrated the feasibility on a human resource database.
