CODEWITHSUNDEEP
pythonpython-3.xscikit-learn
user9185511
user9185511

Reputation: 814

get sklearn.LabelEncoder() mappings after fit_transform

I'm trying to get the mappings of label encoder to figure out which code got each string for a column in my df.

Here is the encoding code:

y[:]=LabelEncoder().fit_transform(y[:])

I would like to get something like this as output:

A:1
B:2
C:3

Thanks for the help!

Upvotes: 5

Views: 3060

Answers (1)

cs95
cs95

Reputation: 403208

You should refrain from in-line initialisation if you want to be able to make use of the mappings or inverse-transformation later.

data = ['A', 'A', 'B', 'C', 'B', 'B'] # `y`

le = LabelEncoder()
mapped = le.fit_transform(data)

mapping = dict(zip(le.classes_, range(1, len(le.classes_)+1)))
print(mapping)
# {'A': 1, 'B': 2, 'C': 3}

Better still, if you want to reverse the encoding, use inverse_transform:

print(le.inverse_transform(mapped))
# ['A' 'A' 'B' 'C' 'B' 'B']

Upvotes: 4

Related Questions