How to Use LabelEncoder in sklearn with Python

Use LabelEncoder from sklearn.preprocessing to convert categorical labels into numeric values. Fit the encoder with fit() or fit_transform() on your label data, then transform labels to numbers with transform().

Syntax

The basic syntax to use LabelEncoder is:

• from sklearn.preprocessing import LabelEncoder: Import the class.
• encoder = LabelEncoder(): Create an encoder object.
• encoder.fit(labels): Learn the unique labels.
• encoded_labels = encoder.transform(labels): Convert labels to numbers.
• Or use encoded_labels = encoder.fit_transform(labels) to fit and transform in one step.

python

from sklearn.preprocessing import LabelEncoder

encoder = LabelEncoder()
encoder.fit(labels)  # Learn unique labels
encoded_labels = encoder.transform(labels)  # Convert labels to numbers

# Or combine:
encoded_labels = encoder.fit_transform(labels)

Example

This example shows how to convert a list of fruit names into numeric labels using LabelEncoder. It fits the encoder and transforms the labels, then prints the original and encoded labels.

python

from sklearn.preprocessing import LabelEncoder

fruits = ['apple', 'banana', 'apple', 'orange', 'banana', 'orange']

encoder = LabelEncoder()
encoded_fruits = encoder.fit_transform(fruits)

print('Original labels:', fruits)
print('Encoded labels:', encoded_fruits)
print('Classes:', encoder.classes_)

Output

Original labels: ['apple', 'banana', 'apple', 'orange', 'banana', 'orange']
Encoded labels: [0 1 0 2 1 2]
Classes: ['apple' 'banana' 'orange']

Common Pitfalls

Common mistakes when using LabelEncoder:

• Trying to encode features instead of labels. LabelEncoder is meant for target labels, not input features.
• Using transform() before fit() causes errors because the encoder doesn't know the classes yet.
• Encoding training and test labels separately can cause inconsistent mappings. Always fit on training labels and transform test labels.
• LabelEncoder only works on 1D arrays or lists of labels, not on multi-dimensional data.

python

from sklearn.preprocessing import LabelEncoder

labels_train = ['cat', 'dog', 'cat']
labels_test = ['dog', 'cat']

encoder = LabelEncoder()
encoder.fit(labels_train)  # Correct: fit on training labels
encoded_train = encoder.transform(labels_train)
encoded_test = encoder.transform(labels_test)  # Use transform, not fit_transform

print(encoded_train)
print(encoded_test)

Output

[0 1 0]
[1 0]

Quick Reference

LabelEncoder Quick Tips:

• Use for encoding target labels, not features.
• Fit once on training labels, then transform all data.
• Use fit_transform() to combine fit and transform.
• Access original classes with encoder.classes_.
• Encoded labels are integers starting from 0.

Key Takeaways

LabelEncoder converts categorical labels into numeric form for ML models.

Always fit the encoder on training labels before transforming.

Use fit_transform() to fit and encode in one step for convenience.

Do not use LabelEncoder on input features; it is for target labels only.

Access encoder.classes_ to see the original label order.