Deep Web Schema Matching Based on Concept-Word and Semantic-Heterogeneous Model
Lijun Chen · 2016
Deep Web contains a vast amount of valuable information, which can be picked and exploited by data integration. A crucial step of Deep Web integration is discovering semantic correspondences over multiple Web databases, which known as schema matching. In this paper, we propose a schema matching method based on concept-word and semantic-heterogeneity for Deep Web. Firstly, the method preprocesses schema through extracting concept-words, discriminates and combines group attributes to convert complex matching into simple matching for improving implement efficiency. Secondly, by introducing instance into semantic-heterogeneous model, the problem of mining semantic-heterogeneous synonymy attributes is resolved by computing, synthetic evaluating, and selecting similarity values of attribute features. Experimental results indicate that compared with evidence theory method, the efficiency and accuracy of our method is improved obviously.