查找每个目标类别中句子中每个单词的出现次数

Question

我有类似的东西。

Sentence                                        Target
We regret to inform you about the result.        1
We are glad to inform you about the result.      2
We would like to inform you about the result.   3
We are surprised to see the result.              4

我想要一个看起来像这样的字数

Word    Target 1    Target 2    Target 2    Target 4
Result     1           1            1           1
Inform     1           1            1           0
Surprised   0           0           0           1

......等等。我该怎么做？

Answer 1

您需要

1. 删除标点符号并小写数据
2. 拆分空白
3. stack创建一个系列
4. groupby Target
5. 找到每个目标的value_counts个字词
6. unstack您所需输出的结果

df.Sentence.str.replace('[^\w\s]', '')\
  .str.lower()\
  .str.split(expand=True)\
  .set_index(df.Target)\
  .stack()\
  .groupby(level=0)\
  .value_counts()\
  .unstack(0, fill_value=0)\
  .add_prefix('Target ')


Target     Target 1  Target 2  Target 3  Target 4
about             1         1         1         0
are               0         1         0         1
glad              0         1         0         0
inform            1         1         1         0
like              0         0         1         0
regret            1         0         0         0
result            1         1         1         1
see               0         0         0         1
surprised         0         0         0         1
the               1         1         1         1
to                1         1         1         1
we                1         1         1         1
would             0         0         1         0
you               1         1         1         0