The proliferation of linked medical data on the Web has increased rapidly in terms of both the number of repositories and the datasets’ sizes. These data have a formal format that is RDF (Resource Description Framework). However, the size and complexity of these medical linked data, leading to an error in modeling, missing concepts as well as missing relationships and inconsistencies in critical clinical applications and biomedical research. Thus, to overcome these issues, this paper aims to describe an approach that uses neural language models for RDF data clustering to identify a thematic view of the clustered entities to extract labels or tags even if there are missing entities or links in the integration of medical RDF datasets. Experiments on a benchmarking dataset give us promising results.
CITATION STYLE
Eddamiri, S., Zemmouri, E., & Benghabrit, A. (2021). Theme identification for linked medical data. In Lecture Notes in Networks and Systems (Vol. 144, pp. 145–157). Springer. https://doi.org/10.1007/978-3-030-53970-2_14
Mendeley helps you to discover research relevant for your work.