This paper describes a method of generating definitions automatically by making the meaning of morphemes explicit, and applies this method to medical neologisms. First, the morphemes that make up the neologisms are automatically recognized and each morpheme is assigned its corresponding meaning. These meanings are then combined through rewrite rules, in order to obtain a noun phrase that expresses the meaning of the neologism. The procedure was tested on a list of 190 medical neologisms extracted from the CCM-2009 corpus (Burdiles, 2012) and obtained 74.34% precision, 64.21% recall, and a 68.9% F-measure.
|Translated title of the contribution||Automatic generation of definitions through explicitation: An application to neologisms in the medical domain|
|Number of pages||11|
|State||Published - Dec 2016|