【版本】 AcroMed 1.1
Laboratory for Linguistics and Computation of Brandeis University and School of Medicine of Tufts University, funded by a grant from the National Library of Medicine.
AcroMed is a computer generated database of biomedical acronyms and the associated long forms extracted from the last year of Medline abstracts (2001). AcroMed is a part of the Medstract project whose goal is to apply natural language processing technologies to extraction of knowledge from biomedical texts. The AcroMed database is searchable by either the acronym or the associated long form. Each acronym is linked to the abstracts in which it was discovered, and the long forms can be submitted directly to PubMed as searches by a single click. AcroMed also attempts to classify each long form by its semantic type.
The main new features :
Improved Sense Grouping.
Prefix incorporation in the Long Forms, using frequency in the database. e.g.: the acronym "Pol II", had the Long Form: "polymerase II", in Acromed 1.0. Now it has the correct form "RNA polymerase II"
Frequency information on the ocurrence of the definition in the abstracts (It is not the occurrence of the string).
Long Form normalization based on frequency of the long form.
contains over 481,531 acronyms.
Summary by 杨宏伟 on 2003-05-07
Last updated by 杨宏伟 on 2003-05-07