Closed gazzola closed 3 years ago
@gazzola First step in NegBio pipeline is extraction of UMLS concepts from the given text. This is being done by MetaMap.
According to MetaMap paper: An overview of MetaMap: historical perspective and recent advances
One of MetaMap's weaknesses is that it can be applied only to English text. MetaMap's English-centric nature is evident throughout its implementation, not just in its lexical and syntactic algorithms.
So if you want to build NegBio pipeline, first thing is you would need to replace MetaMap by an alternate concept extraction library. After that you would need parse tree and dependency tree creation specific to Portuguese-BR language. And then in NegBio, negation/uncertain patterns also would be needed specific to Portuguese-BR language
Does anyone know any version adapted for Brazilian Portuguese?