Download & Install

http://wordnet.princeton.edu/wordnet/download/current-version/

파이썬에서는 nltk 라이브러리를 설치하면 기본적으로 설치되어 있다.



NLTK Module Usage

Import python nltk module:

1
>>> from nltk.corpus import wordnet

Print coordinate terms:

(Coordinate terms are nouns or verbs that have the same hypernym.)

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
>>> wordnet.synsets("lunch")
[Synset('lunch.n.01'), Synset('lunch.v.01'), Synset('lunch.v.02')]
>>> lunch_n = wordnet.synset("lunch.n.01")
>>> lunch_n.hypernyms()
[Synset('meal.n.01')]
>>> lunch_n.hypernyms()[0].hyponyms()
[Synset('banquet.n.02'), Synset('bite.n.04'), Synset('breakfast.n.01'),
Synset('brunch.n.01'), Synset('buffet.n.02'), Synset('dinner.n.01'),
Synset('lunch.n.01'), Synset('mess.n.04'), Synset('nosh-up.n.01'),
Synset('picnic.n.03'), Synset('ploughman´s_lunch.n.01'), Synset('potluck.n.01'),
Synset('refection.n.01'), Synset('square_meal.n.01'), Synset('supper.n.01'),
Synset('tea.n.02')]
>>> lunch_n.lemmas()
[Lemma('lunch.n.01.lunch'), Lemma('lunch.n.01.luncheon'),
Lemma('lunch.n.01.tiffin'), Lemma('lunch.n.01.dejeuner')]
1
2
3
4
5
6
7
8
9
10
>>> lunch_v = wordnet.synset("lunch.v.01")
>>> lunch_v.hypernyms()
[Synset('eat.v.02')]
>>> lunch_v.hypernyms()[0].hypernyms()
[Synset('consume.v.02')]
>>> lunch_v.hypernyms()[0].hyponyms()
[Synset('break_bread.v.01'), Synset('breakfast.v.01'), Synset('brunch.v.01'),
Synset('dine.v.01'), Synset('eat.v.01'), Synset('eat_in.v.01'), Synset('eat_out.v.01'),
Synset('feast.v.01'), Synset('gorge.v.01'), Synset('lunch.v.01'), Synset('mess.v.01'),
Synset('nosh.v.01'), Synset('picnic.v.01'), Synset('take_out.v.12'), Synset('victual.v.03')]

Word relations:
(Antonyms, Derivationally_Related_Forms and Pertainyms are defined only over Lemmas.)

1
2
3
4
5
6
7
8
9
10
11
>>> wordnet.synsets("vocal")
[Synset('vocal_music.n.01'), Synset('song.n.01'), Synset('vocal.a.01'), Synset('vocal.a.02'), Synset('outspoken.s.01'), Synset('vocal.s.04')]
>>> vocal = wordnet.synset("vocal.a.01")
>>> vocal.lemmas()
[Lemma('vocal.a.01.vocal')]
>>> vocal.lemmas()[0].derivationally_related_forms()
[Lemma('vocalize.v.02.vocalize')]
>>> vocal.lemmas()[0].pertainyms()
[Lemma('voice.n.02.voice')]
>>> vocal.lemmas()[0].antonyms()
[Lemma('instrumental.a.01.instrumental')]

Extract lemma:

(Lemma is the base form(headword; lowercase) for a certain word in dictionary.)

1
2
3
>>> from nltk.stem.wordnet import WordNetLemmatizer
>>> WordNetLemmatizer().lemmatize('having','v')
'have'



Synset types

  • n: Noun
  • v: Verb
  • a: Adjective
  • s: Adjective Satellite
  • r: Adverb



Glossary of terms

http://wordnet.princeton.edu/wordnet/man/wngloss.7WN.html#sect4



Multilingual WordNet

http://globalwordnet.org/



References



SENSProgramming Note

티스토리툴바