如何在WordNet中找到两个同义词集之间的语义关系

Question

我正在使用WordNet 3.0。我想知道如何在WordNet层次结构中的两个同义词集之间找到语义关系。例如，给定两个词作为输入，我想找到它们之间的关系，即它们是否是同义词，下位词-上位词等。

是否有python或perl模块来实现这一目标？

Answer 1

NLTK是最常用的Python NLP库。所描述的操作将类似于：

from nltk.corpus import wordnet

house = wordnet.synset('house.n.01')
station = wordnet.synset('station.n.01')

然后您可以通过这种方式使用Synset类的方法：

taxonomy_distance = house.shortest_path_distance(station)

确定它们是否是同义词：

common_lemmas = len(set(house.lemma_names).intersection(set(station.lemma_names)))

这将返回两组中的通用引理数。

您还可以使用像Neo4j这样的图形数据库来加载wordnet数据集，并寻找它们的节点之间的最短路径，这是一个讨论过的问题here。

Answer 2

使用Python NLTK：

>>> from nltk.corpus import wordnet as wn
>>> 
>>> synset1 = wn.synset('adornment.n.01')
>>> synset2 = wn.synset('jewelry.n.01')
>>> 
>>> synset1.lowest_common_hypernyms(synset2)
[Synset('adornment.n.01')]
>>> 
>>> synset1.hyponyms()
[Synset('frill.n.03'), Synset('rosette.n.01'), Synset('frontlet.n.01'), Synset('cordon.n.03'), Synset('fob.n.02'), Synset('beauty_spot.n.01'), Synset('sequin.n.01'), Synset('war_paint.n.01'), Synset('boutonniere.n.01'), Synset('trimming.n.02'), Synset('pendant.n.01'), Synset('pompon.n.01'), Synset('band.n.04'), Synset('bangle.n.02'), Synset('jewelry.n.01'), Synset('epaulet.n.01'), Synset('circlet.n.02'), Synset('frog.n.03'), Synset('tassel.n.01'), Synset('plume.n.02'), Synset('pectoral.n.02')]
>>> synset2.hypernyms()
[Synset('adornment.n.01')]

获得同义词：

>>> def lemmas_in_all_synsets(keyword):
...     lemmas=[]
...     for synset in wn.synsets(keyword):
...         for lemma in synset.lemmas:
...             lemmas.append(lemma)
...     return lemmas
...
>>> lemmas_in_all_synsets('station')
[Lemma('station.n.01.station'), Lemma('place.n.10.place'), Lemma('place.n.10.sta
tion'), Lemma('station.n.03.station'), Lemma('post.n.01.post'), Lemma('post.n.01
.station'), Lemma('station.n.05.station'), Lemma('station.v.01.station'), Lemma(
'station.v.01.post'), Lemma('station.v.01.send'), Lemma('station.v.01.place')]

Answer 3

def synonyms(x):
    from nltk.corpus import wordnet as wn
    synonyms_x = []
    for syn in wn.synsets(x):
        for l in syn.lemmas():
            synonyms_x.append(l.name())
    #to remove duplicates from the list
    synonyms_x = list(dict.fromkeys(synonyms_x)) 
    return synonyms_x