How to use MultiLabelBinarizer for Multilabel classification?

Question

I am trying to do multilabel classification. But I am really stuck at data preprocessing. My target data in in a separate file. The target data looks like this

   Id              Tag
0   1             data
1   4               c#
2   4         winforms
3   4  type-conversion
4   4          decimal

I am trying to use MultiLabelBinarizer to preprocess the data. At the end of it, I want it to look something like this -

ID	data	c#	winforms	type-conversion	decimal
1	1	0	0	0	0
4	0	1	1	1	1

This is the code that I am using

import pandas as pd
from sklearn.preprocessing import MultiLabelBinarizer

question_tags = pd.read_csv("./archive/question_tags.csv")
print(question_tags.head())
mlb = MultiLabelBinarizer()
print(mlb.fit_transform(question_tags))

This is the output I am getting.

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

What am I doing wrong?

Answer 1

import pandas as pd
from sklearn.preprocessing import MultiLabelBinarizer

data = pd.DataFrame({
    'id': [1, 4, 4, 4, 4],
    'tag': ['data', 'c#', 'winforms', 'type-conversion', 'decimal']
})

gr = data.groupby('id')['tag'].apply(list).reset_index()
print(gr)
'''
   id                                       tag
0   1                                    [data]
1   4  [c#, winforms, type-conversion, decimal]
'''
mlb = MultiLabelBinarizer()
tag_binarized = mlb.fit_transform(gr['tag'])
tags_df = pd.DataFrame(tag_binarized,columns = mlb.classes_)
print(tags_df)
'''
  c#  data  decimal  type-conversion  winforms
0   0     1        0                0         0
1   1     0        1                1         1
'''
res = pd.concat([gr['id'],tags_df],axis=1)
print(res)
'''
  id  c#  data  decimal  type-conversion  winforms
0   1   0     1        0                0         0
1   4   1     0        1                1         1
'''

Answer 2

You are not doing anything wrong. MultiLabelBinarizer(), as most other sklearn stuff, returns numpy arrays. In this case, the underlying data looks identical to your expected output, sans the ID and Tag names.

Use pd.crosstab instead:

pd.crosstab(df['Id'], df['Tag'])