如何在字符级别对句子进行单热编码?

问题描述

我想将一个句子转换为一个单热向量数组.这些向量将是字母表的 one-hot 表示.它看起来像下面这样:

I would like to convert a sentence to an array of one-hot vector.These vector would be the one-hot representation of the alphabet.It would look like the following:

"hello" # h=7, e=4 l=11 o=14

会变成

[[0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0]
 [0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0]
 [0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0]
 [0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0]
 [0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0]]

不幸的是，来自 sklearn 的 OneHotEncoder 没有作为输入字符串.

Unfortunately OneHotEncoder from sklearn does not take as input string.

推荐答案

只需将传递的字符串中的字母与给定的字母表进行比较:

Just compare the letters in your passed string to a given alphabet:

def string_vectorizer(strng, alphabet=string.ascii_lowercase):
    vector = [[0 if char != letter else 1 for char in alphabet]
                  for letter in strng]
    return vector

请注意，使用自定义字母表(例如defbcazk"，列将在每个元素出现在原始列表中时进行排序).

Note that, with a custom alphabet (e.g. "defbcazk", the columns will be ordered as each element appears in the original list).

string_vectorizer('hello')的输出:

[[0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0],
 [0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0],
 [0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0],
 [0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0],
 [0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0]]

这篇关于如何在字符级别对句子进行单热编码?的文章就介绍到这了，希望我们推荐的答案对大家有所帮助，也希望大家多多支持！