Theprimary method for scientific communication is in the form ofpublished scientific articles and theses and the use of naturallanguage combined with domain-specific terminology. As such, theycontain unstructured data.Giventhe unquestionable usefulness of data extraction from unstructuredliterature, we aim to show how this can be achieved for thediscipline of chemistry. The highly formulaic style of writing mostchemists adopt make their contributions well suited tohigh-throughput Natural Language Processing (NLP) approaches. Usingchemical synthesis procedures as an exemplar, we presentChemicalTagger. ChemicalTagger is a tool that combines chemicalentity recognisers such as OSCAR with tokenisers,  part-of-speechtaggers and shallow parsing tools to produce a formal structure ofreactions.Thisextracted data can then be expressed in RDF. This allows  for thegeneration of highly informative visualisations, such as visualdocument summaries, structured querying and further enrichment can beprovided by linking with domain specific ontologies.

ChemicalTagger:
A tool for semantic text-mining in chemistry

240th National Meeting (2010)

The official archive and repository for all ACS Meetings & Expositions abstracts and select posters or presentations.

ChemicalTagger: A tool for semantic text-mining in chemistry