Semantic Mediation of Metadata for Marine Geochemical Data Integration

With the evolution of sea floor exploration technologies, marine samples and the corresponding geochemical data increased dramatically. The heterogeneity of the data in terms of syntactic and semantic structures becomes an obstacle to knowledge integration, transmission and sharing. In order to solve the problems of long-lived data collection and integration, this paper proposes a semi-automatic method based on a semantic extraction and semantic matching technique. We develop semantic extract and semantic matching tools, which are named Metadata Extraction and Metadata Mapping respectively. Metadata Extraction tool could extract the metadata of users' datasets. Metadata Mapping tool could establish semantic mappings semi-automatically. Moreover, hierarchies of users' metadata are added automatically according to the semantic mapping. The method is applied to marine geochemical data from Cruises of Chinese International Deep Seabed Resource Survey. Experimental results show that our approach not only reduce the workload of data integration, but also immensely eliminate the problem of the semantic heterogeneity.