Semantic Knowledge Base Construction from Domain-Specified Metadata

Existing semantic knowledge bases such as WordNet and Yago contain the information of relations between entities. Although they are not domain-specific, they contain limited information and have limitations on the usage scopes and domains. For example, they do not hold the information about the domain-specific commonsense relations between concepts like “horse” and “farm” or “ocean” and “atmosphere” which intuitively have close relations on semantics in the domains of image description or earth observation data description. Such semantic knowledge are useful in the corresponding domains for various applications such as annotation, recommendation, search, suggestion and so on. Because in data collections, metadata which is used to describe data is widespread in various domains, in this paper we propose an approach to collect this kind of relations and construct knowledge bases for specific domains by mining knowledge of global structure and internal association in the metadata of data collections.