An Approach to Database Schema Matching Based on Entity Classification

Schema matching plays a key role in many application domains,such as data integration,data warehouse, and information share and exchange on computer network. Currently,approaches of automatic schema matching cannot solve matching issue under the circumstance of complex schema well. This paper introduces an approach based on entity classification in the domain of relation schema. It divides entities into different categories (sub-schema) using Na ive Bayes Learning ,and then matches schema elements between the sub-schemas with the same category. It can effectively improve matching results, reduce the number of element-to-element comparisons and save user efforts under the circumstance of complex schema.