EDFund- Elsevier funding entity linking dataset
Automatic extraction of funding information from academic articles adds significant value to industry and research communities, such as tracking research outcomes by funding organizations, profiling researchers and universities based on the received funding, and supporting open access policies. This dataset collected by Elsevier in 2019. Sentences talking about the funding information are extracted from 30000 scientific articles. Name of organisation which financially supported the research is highlighted and linked to Crossref Taxonomy. Each document is annotated by two human and harmonized by an independent SME. For more information please read our paper "Find the Funding: Entity Linking with Incomplete Funding Knowledge Bases" presented in COLING2022. In addition to EDFund, we also release ELFund. This dataset only includes the sentences which have been classified as positive by our sentence classifier, as explained in the aforementioned paper. ELFund: #Articles: 53864 #Links: 177111 NILs%: 19.23% EDFund: #Articles: 53243 #Links: 172031 NILs%: 18.95% - We thank Ramadurai Petchiappan and Georgios Cheirmpos for their assistance in preparing the 2nd version of the datasets.
Steps to reproduce
Each document is annotated by two human and harmonized by an independent SME.
Additional metadata for Elsevier datasets
|Date the data was collected||2019-07-01T00:00:00.000Z|