LabelEncoder

The OrdinalEncoder and OneHotEncoder are usually used to encode features (the X variable).
But the target (y variable) can also be categorical.
Let's look at the example:


              123456789
            
import pandas as pd

# Load the data and assign X, y variables
df = pd.read_csv('https://codefinity-content-media.s3.eu-west-1.amazonaws.com/a65bbc96-309e-4df9-a790-a1eb8c815a1c/adult_edu.csv')
y = df['income'] # Income is a target in this dataset
X = df.drop('income', axis=1)

print(y)
print('All values: ', y.unique())

The LabelEncoder is used to encode the target, regardless of whether it is nominal or ordinal.

ML models do not consider the order of the target, allowing it to be encoded as any numerical values. LabelEncoder encodes the target to numbers 0, 1, ...


              1234567891011121314
            
import pandas as pd
from sklearn.preprocessing import LabelEncoder

# Load the data and assign X, y variables
df = pd.read_csv('https://codefinity-content-media.s3.eu-west-1.amazonaws.com/a65bbc96-309e-4df9-a790-a1eb8c815a1c/adult_edu.csv')
y = df['income'] # Income is a target in this dataset
X = df.drop('income', axis=1)
# Initialize a LabelEncoder object and encode the y variable
label_enc = LabelEncoder()
y = label_enc.fit_transform(y)
print(y)
# Decode the y variable back
y_decoded = label_enc.inverse_transform(y)
print(y_decoded)

The code above encodes the target using LabelEncoder and then uses the .inverse_transform() method to convert it back to the original representation.

Note
Since the LabelEncoder is used to transform target (y), which is usually a single column, it works well with pandas Series, unlike the OrdinalEncoder.
So we can just pass the y variable to the .fit_transform() method.

Everything was clear?

Thanks for your feedback!

Section 2. Chapter 7