Ontology-Enabled Metadata Schema Generator: The Design Approach
Jian Qin, Xiaozhong Liu, Miao Chen · 2015
Metadata standards are important for normalizing descriptions of publications and research data and for information discovery and use. Large, complex metadata standards, however, can complicate the creation, sharing, and maintenance of metadata and incur high costs for metadata operations, especially in the domain of scientific data (Qin et al., 2010; Qin Ball, & Greenberg, 2012; Qin & Li, 2013). One strategy to solve the problems of large, complex metadata standards is to break them into independent modules to allow for reuse of elements and maximal possibility of automation. To implement this strategy, we need a metadata infrastructure that contains elements, vocabularies, and other metadata artifacts and that is easy to use. This short paper describes the design approach to an ontology-enabled metadata schema generator as part of the metadata infrastructure. Elements in metadata standards in the scientific data domain tend to follow a pattern that a small number of (super-) general elements co-occur in a large number of standards and those co-occurred in 2-4 standards tend to be field-general. Even though semantically same elements co-occurred across different standards, they often varied in singular-plural forms, capitalization, or complete different words (Qin & Li, 2013). These inconsistencies and varying naming