SMGM：一种基于模式结构和已有匹配知识的模式匹配模型

模式匹配是确定模式间语义匹配关系的技术，它在许多应用中起着重要的作用，如数据集成中异构模式信息整合、本体知识映射、电子商务中消息映射等。针对已有模式匹配方法的局限性，本着最大限度地减少人工干预使模式匹配自动化的原则，本文提出一种利用模式结构信息和已有匹配知识的模式匹配模型SMGM。它借鉴神经网络元间影响作用过程实现语义匹配推理；通过重用已有匹配知识，补充、精化匹配知识，自动缩减不确定阈值区间；并给出一种自适应式迭代挖掘求精已有匹配知识的自学习型模式匹配模型。实验表明：SMGM模型切实可行。

SMGM: One Schema Matching Model Based on Schema Structures and Known Matching Knowledge

Schema matching is the task of finding semantic correspondences between elements of two schemas. It is critical in many applications, such as data integration, data warehouse loading and XML message mapping, etc. Against the limitations of existed schema matching methods,with the aim of reducing the amount of user effort as much as possible to automatic schema matching, based on the schema structure information and known matching knowledge, we propose a novel approach to schema matching method called SMGM. It imitates the influence procedure between neurons to realize the semantic matching reasoning. By reusing the known matching knowledge to supplement and dive the matching knowledge and curtail the uncertain threshold interval automatically, and presented a self learning schema matching model which can mine and dive the known matching knowledge adaptively and iterately. The result of our experiment shows that the SMGM is feasible.