Abstract:

Due to autonomy and heterogeneity data sources, uncertainty is an inherent character of schema matching.  In order to improve the performance of schema matching, an uncertainty matching approach based on evidence theory is proposed. Firstly, the schema space is divided into several schema subspaces according to attributes types. Secondly, different matchers are viewed as different sources of evidence, and mass distributions are defined on the basis of the match results from these matchers. Thirdly, an improved evidence theory is used to automatically combine multiple matchers, which reduces human involvement and solves the situations with high conflict results from different matchers. Finally the mapping is generated by the improved KuhnMunkres algorithm. The experiments show that the proposed method is highly accurate and effective.

Key words: schema matching;uncertainty;matchers;evidence theory