Label encoding in ML Python

Introduction

Label encoding changes words or categories into numbers so computers can understand them.

When you have categories like colors or types and want to use them in a machine learning model.

When you need to convert text labels into numbers for classification tasks.

When preparing data for algorithms that only accept numbers.

When you want a simple way to turn categories into numbers without creating many new columns.

Syntax

ML Python

from sklearn.preprocessing import LabelEncoder

le = LabelEncoder()
encoded_labels = le.fit_transform(list_of_labels)

fit_transform learns the categories and converts them to numbers in one step.

The numbers start from 0 and go up to the number of categories minus one.

Examples

This example converts three color names into numbers: blue=0, green=1, red=2.

ML Python

from sklearn.preprocessing import LabelEncoder

labels = ['red', 'green', 'blue']
le = LabelEncoder()
encoded = le.fit_transform(labels)
print(encoded)

Repeated categories get the same number: bird=0, cat=1, dog=2.

ML Python

from sklearn.preprocessing import LabelEncoder

le = LabelEncoder()
labels = ['cat', 'dog', 'cat', 'bird']
encoded = le.fit_transform(labels)
print(encoded)

Sample Model

This program turns fruit names into numbers so a computer can use them. It also shows which number matches each fruit.

ML Python

from sklearn.preprocessing import LabelEncoder

# List of fruit names
fruits = ['apple', 'banana', 'cherry', 'banana', 'apple', 'cherry']

# Create LabelEncoder object
le = LabelEncoder()

# Fit and transform the fruit list
encoded_fruits = le.fit_transform(fruits)

print('Original labels:', fruits)
print('Encoded labels:', encoded_fruits)

# Show the mapping from labels to numbers
for label, code in zip(le.classes_, range(len(le.classes_))):
    print(f"'{label}' is encoded as {code}")

Important Notes

Label encoding assigns numbers based on alphabetical order of categories.

It is best for categories without order or when the model can handle numeric labels properly.

For categories with no natural order, consider one-hot encoding to avoid implying order.

Summary

Label encoding changes categories into numbers so machines can understand them.

It is simple and useful for many classification problems.

Always check if label encoding fits your data type and model needs.