Konlpy¶

한글모듈

1 Tagging¶

In [1]:

from konlpy.tag import Twitter
twitter = Twitter()
twitter.pos('파이썬을 활용한 자연어 수업자료')

Out[1]:

[('파이썬', 'Noun'),
 ('을', 'Josa'),
 ('활용한', 'Verb'),
 ('자연어', 'Noun'),
 ('수업', 'Noun'),
 ('자료', 'Noun')]

2 Stemming¶

In [2]:

text = "워런 버핏은 삼성전자가 아닌 애플주식을 왜 샀을까"

In [3]:

%%time
print(twitter.pos(text, stem=True))

[('워런', 'Noun'), ('버핏', 'Noun'), ('은', 'Josa'), ('삼성', 'Noun'), ('전자', 'Noun'), ('가', 'Josa'), ('아니다', 'Adjective'), ('애플', 'Noun'), ('주식', 'Noun'), ('을', 'Josa'), ('왜', 'Noun'), ('사다', 'Verb')]
CPU times: user 821 ms, sys: 34 ms, total: 855 ms
Wall time: 564 ms

In [4]:

%%time
print(twitter.pos(text))

[('워런', 'Noun'), ('버핏', 'Noun'), ('은', 'Josa'), ('삼성', 'Noun'), ('전자', 'Noun'), ('가', 'Josa'), ('아닌', 'Adjective'), ('애플', 'Noun'), ('주식', 'Noun'), ('을', 'Josa'), ('왜', 'Noun'), ('샀', 'Verb'), ('을까', 'Eomi')]
CPU times: user 48.4 ms, sys: 26 µs, total: 48.4 ms
Wall time: 67 ms

3 꼬꼬마, 한나눔¶

개별 모듈에 따라 다른결과를 출력

In [5]:

%%time
from konlpy.tag import Kkma
kkma = Kkma()
print(kkma.pos(text))

[('워', 'UN'), ('런', 'NNG'), ('버핏', 'UN'), ('은', 'JX'), ('삼성전자', 'NNG'), ('가', 'JKS'), ('아니', 'VV'), ('ㄴ', 'ETD'), ('애플', 'NNP'), ('주식', 'NNG'), ('을', 'JKO'), ('왜', 'MAG'), ('사', 'VV'), ('었', 'EPT'), ('을까', 'EFQ')]
CPU times: user 21.6 s, sys: 229 ms, total: 21.8 s
Wall time: 16.1 s

In [6]:

%%time
from konlpy.tag import Hannanum
han = Hannanum()
print(han.pos(text))

[('워런', 'N'), ('버핏', 'N'), ('은', 'J'), ('삼성전자', 'N'), ('가', 'J'), ('아니', 'P'), ('ㄴ', 'E'), ('애플주식', 'N'), ('을', 'J'), ('왜', 'M'), ('사', 'P'), ('아ㄹ까', 'E')]
CPU times: user 8.6 s, sys: 77 ms, total: 8.68 s
Wall time: 5.29 s