NER干扰REGEXNER [英] NER interfere with REGEXNER

问题描述

我使用regexner查找不在Stanford NLP的默认集合中的命名实体，并且工作正常.但是，当我添加ner注释器时，它将使用默认标签注释与我的正则表达式匹配的标记.如何覆盖默认注释?

I use regexner to find named entities that are not in the default set of Stanford NLP and it works fine. However, when I add ner annotator, it annotates tokens that match my regular expression with default tags. How can I overwrite default annotations?

def createNLPPipelineRegex(): StanfordCoreNLP = {
     val props = new Properties()
     props.put("regexner.mapping", "regex.txt")
     props.put("annotators", "tokenize, ssplit, regexner, pos, lemma, ner")
     props.put("tokenize.options", "untokenizable=noneKeep,normalizeParentheses=false")
     new StanfordCoreNLP(props)

}

推荐答案

如果您在ner注释器后添加regexner，它将起作用:

If you add regexner after the ner annotator it should work:

props.put("annotators", "tokenize, ssplit, pos, lemma, ner, regexner")

这篇关于NER干扰REGEXNER的文章就介绍到这了，希望我们推荐的答案对大家有所帮助，也希望大家多多支持IT屋！