개체명인식

[8-1] NLTK를 이용한 개체명 인식

# -*- coding:utf-8 -*-
import nltk

nltk.download('punkt')
nltk.download('words')
nltk.download('averaged_perceptron_tagger')
nltk.download('maxent_ne_chunker')

sentence = "Prime Minister Boris Johnson had previously said the UK would leave by 31 October."

tokens = nltk.word_tokenize(sentence)
print(tokens)
['Prime', 'Minister', 'Boris', 'Johnson', 'had', 'previously', 'said', 'the', 'UK', 'would', 'leave', 'by', '31', 'October', '.']

tagged = nltk.pos_tag(tokens)
print(tagged)
[('Prime', 'NNP'), ('Minister', 'NNP'), ('Boris', 'NNP'), ('Johnson', 'NNP'), ('had', 'VBD'), ('previously', 'RB'), ('said', 'VBD'), ('the', 'DT'), ('UK', 'NNP'), ('would', 'MD'), ('leave', 'VB'), ('by', 'IN'), ('31', 'CD'), ('October', 'NNP'), ('.', '.')]

entities = nltk.chunk.ne_chunk(tagged)
print(entities)

(S
  Prime/NNP
  Minister/NNP
  (PERSON Boris/NNP Johnson/NNP)
  had/VBD
  previously/RB
  said/VBD
  the/DT
  (ORGANIZATION UK/NNP)
  would/MD
  leave/VB
  by/IN
  31/CD
  October/NNP
  ./.)