Geek Answers Handbook

Is there a POS tag capable of parsing "When did Nixon die?" correctly?

I use the Stanford POS tester to analyze questions. This is one of the best POS taggers available in my experience, but a simple question like "When did Nixon die?" incorrectly annotated as

When_WRB did_VBD Nixon_NNP die_NN

The problem is "die_NN". Obviously, "death" is not a noun in this case, but a verb. This happens with a lot of questions. My questions:

  • Can I customize the shorthand of Stanford POS tester to give the correct annotation for this sentence? Or, alternatively, let him create several annotations, of which one is correct?
  • Are there any other Java POS POS tags that can correctly parse such issues?
+7
java stanford-nlp pos-tagger
source share
2 answers

As stated in the comments, "The Tagger Stanford POS Tester is currently getting it right."

Lol ... I will take generosity if he does not want it.

0
source share

The part of the library that I wrote seems to do this. It identifies what parts the words belong to (NOUN, ADVERB, etc.), and correctly understands:

import cortecx.construction.cxsession as cxsession from cortecx.construction.materials import Parser from pprint import pprint cxsession.Session().start() parser = Parser() parser.include('chunks') pprint(parseronie('When did Nixon die?'))

and here is the JSON output:

 {'analysis': [{'when': {'chunk': 'B-ADVP'}}, {'did': {'chunk': 'O'}}, {'nixon': {'chunk': 'B-NP'}}, {'die': {'chunk': 'B-ADVP'}}, {'?': {'chunk': 'O'}}]}

The library can also do POS, NER, etc. .... if you want to play with it yourself, it is worth checking out at https://www.cerybra.com/

0
source share

More articles:


All Articles