A method on lexical disambiguation in distributed heterogeneous autonomous database

Abstract

Schema matching is a fundamental process in many database applications and its quality impacts the performance of such applications. However, finding semantic similarities among schema is a challenging issue in developing full automatic schema matching of distributed heterogeneous autonomous database. Lexical ambiguity is a pervasive phenomenon among these database schemas that prevent fully automatically matching process. Consequently, it is often hard to determine the correct semantic similarities between schemas without full understanding intend meaning associate with schema element name. Lexical ambiguity leads to semantic conflicts (homonym and synonym) on schema matching. Several approaches of schema matching have been proposed. However, these approaches are still limited as they does not disambiguate schema concept, while it just evaluates the possible meanings. Therefore, the aim of this paper is to propose a disambiguation method to eliminate ambiguity. To avoid semantic conflicts, a further enhance the results of schema matching is required to improve the usability of heterogeneous data sources. As work progress, future effort will develop full automatic schema matching and integration.