An algorithm for knowledge integration and refinement

Knowledge base is the foundation of intelligent system. It is very important to insure the consistency and non-redundancy of knowledge in knowledge base. The redundant, inclusive and incompatible knowledge must be processed in knowledge-integration due to variety of knowledge source. In this paper, we research the incompatible knowledge elimination approach in knowledge-integration based on rough set theory, and present a new knowledge integration algorithm KIRRS, which is effective to improve the efficiency of knowledge-integration.