Semi-Automatic Generating Semantic Markup Webpage from Structured Data with Semantic Matching

Abstract

Along with the development of the knowledge graph, it required more significant structured data to build it. Publishing massive structured data, such as relational databases, require substantial effort. Therefore, this work pro-poses a framework for semi-automatic generating structured data and creating a semantic matching algorithm for the table’s attributes and schema.org’s properties. The algorithm uses Wu Palmer Similarity (WUP) and WordNet as semantic similarity measurements. Although the obtained scores are still pretty fair, the whole framework runs well at 0.4285 for WUP+k and 0,3833 for WUP.