/：'Counter'和'int'的不受支持的操作数类型-单词袋

Question

I'm attempting to plot my bag of word for selecting a suitable number of words to use to train my regression algorithm. When I try to plot it though, I get an error.

This is my code for making the bag of words:

df['BOW'] = df.Review2.str.split().apply(Counter)
df['BOW'].head()

This is the code for the plotting (I use seaborn ):

sns.distplot(df['BOW'].sum())

and this is the error:

TypeError: unsupported operand type(s) for /: 'Counter' and 'int'

this is how my bag of words looks like:

0 {'good': 2, 'need': 1, 'change': 1, 'virgils':...

1 {'new': 3, 'favorite': 2, 'give': 1, 'delightf...

2 {'red': 3, 'sauce': 2, 'favorite': 1, 'enjoy':...

3 {'quality': 1, 'fantastic': 1, '1800s': 1, '21...

4 {'red': 1, 'first': 1, 'time': 1, 'try': 1, 'l...

Any insight is appreciated!

Answer 1

You're generating Counter objects.

>>> Counter('some words here are some repeated words'.split())
Counter({'words': 2, 'repeated': 1, 'some': 2, 'here': 1, 'are': 1})

While you can sum() counter objects, ...

>>> Counter('aaabb') + Counter('abc')
Counter({'a': 4, 'b': 3, 'c': 1})

... you can't divide them to calculate average:

>>> (Counter('aaabb') + Counter('abc')) / 2
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
TypeError: unsupported operand type(s) for /: 'Counter' and 'int'

That's what your plotting library seems to be doing? Maybe you want to remove that column from the graph....