Linked Data Profiling Identifying the Domain of Datasets Based on Data Content and Metadata

Since the beginning of the Linked Open Data initiative, the number of published open datasets has gradually increased, but the datasets often do not contain description about content such as the dataset domain (e.g., medicine, cancer), when this information is available, it is usually coarse-grained e.g. organic-edunet contains the metadata about a collection of learning objects exposed through the Organic.Edunet portal, but it is classified as Life science. In this work we propose approaches that will provide a detailed description of existing datasets as well as linking assistance when publishing new datasets by generating detailed descriptions of the publishers dataset.