SDD-matcher: a semantic-driven data matching framework

摘要

A generic semantic-driven data matching framework (SDD-Matcher) has been designed and developed for matching data objects across organizations. It contains matching algorithms at three different levels: string, lexical and graph. The level of graph is also called ontological or conceptual level. Each SDD-Matcher matching strategy at least contains a graph matching algorithm. Besides, we can freely choose the matching algorithms at these three levels when composing a matching strategy. A matching strategy can as well be a composition of several matching strategies. In this article, the strategies of LexMA, OntoGraM, C-FOAM and GRASIM are elaborated. We focus on a full engineering cycle of SDD-Matcher, naming use casestudy, design and formalization, an evaluation method, implementation and evaluation. We use real case data for the evaluation, which has been validated by large enterprises. The recorded highest users' satisfaction rate is 78%. With the proposed iterations in this full engineering cycle, the satisfaction rate can potentially increase by updating the parameters in the strategies based on users' feedbacks.