如何从nltk pos_标记获取标记集？

Question

我试图从nltk pos_标记中获取完整标记，但我找不到使用nltk的简单方法。例如，使用tagsets='universal'

from nltk.tokenize import word_tokenize

def nltk_pos(text):
    token = word_tokenize(text)
    return (nltk.pos_tag(token)[0])[1]

nltk_pos('home')
output: 'NN'
expected output: 'NOUN'

Answer 1

我在为我写的一篇论文做NLP分析时也遇到了同样的问题。我必须使用这样的映射函数：

import nltk
from nltk.tokenize import word_tokenize

def get_full_tag_pos(pos_tag):
    tag_dict = {"J": "ADJ",
                "N": "NOUN",
                "V": "VERB",
                "R": "ADV"}
    # assuming pos_tag comes in as capital letters i.e. 'JJR' or 'NN'
    return tag_dict.get(pos_tag[0], 'NOUN')

# example
words = word_tokenize(text)
words_pos = nltk.pos_tag(words)
full_tag_words_pos = [word_pos[0] + "/" + get_full_tag_pos(word_pos[1]) for word_pos in words_pos]