institute of bioinformatics and systems biology / mips

Font size »A . A+ . A++ .

DefineTHAT - automated definition extraction

Manual construction of dictionaries within a scientific domain is a time-consuming and expensive task. Furthermore they are rarely up-to-date since new concepts emerge quickly. With the exponentially growing number of scientific publications, manual maintenance of such lexicon resources is impractical. To solve these problems for the biomedical domain, we developed an automatic definition extraction system based on text mining of abstracts and full-text articles indexed in Pubmed MEDLINE and Pubmed Central. Overall, we extracted approximately 11.1 million sentences containing definitions from currently more than 120 million sentences. Our resource can be searched for any biomedical or non- biomedical term mentioned within the subject part of those sentences. In order to provide access, we developed an Android App, called DefineTHAT, and a web-based client to retrieve definitions from our database, both freely accessible.


Here you can download the android application. ANDROID APP

The web client is accessible here: WEB CLIENT