Automatic extraction of abbreviation definitions based on general texts

Abstract

The study of abbreviation identifications mostly is limited to the biomedical literature. The wide use of abbreviations in general texts, including web data and newswire data, requires us to process and extract the abbreviation definition. In this paper, we propose an abbreviation definition identification algorithm, which employs a variety of rules and incorporates shallow parsing of the text to identify the most probable abbreviation definition from general texts. The performance of our system was tested with data set provided by 2012 NIST1 TAC-KBP2, obtaining a performance of 94.2% recall and 95.5% precision.