Label encoding in ML Python - Practice Problems & Coding Challenges

Challenge - 5 Problems

❓ Predict Output

intermediate

Output of label encoding with unseen category

What will be the output of the following code snippet when label encoding is applied to a test set containing an unseen category?

ML Python

from sklearn.preprocessing import LabelEncoder

train_labels = ['cat', 'dog', 'fish']
test_labels = ['dog', 'cat', 'bird']

le = LabelEncoder()
le.fit(train_labels)
encoded_test = le.transform(test_labels)
print(encoded_test.tolist())

A [1, 0, 0]

B Raises a ValueError due to unseen label 'bird'

C [1, 0, 2]

D [2, 1, 0]

💡 Hint

Think about what happens when the label encoder sees a label it did not learn during fitting.

LabelEncoder cannot transform labels that were not seen during fitting. The unseen label 'bird' causes a ValueError.

🧠 Conceptual

intermediate

Purpose of label encoding in machine learning

Why do we use label encoding for categorical data in machine learning?

A To create new features by combining existing ones

B To normalize numeric features between 0 and 1

C To reduce the size of the dataset by removing categories

D To convert categorical labels into numeric form so algorithms can process them

💡 Hint

Machine learning models usually require numbers, not words.

Label encoding converts categories into numbers so models can understand and process the data.

❓ Metrics

advanced

Effect of label encoding on model accuracy

You have a classification model trained on label encoded target labels. If the label encoding mapping changes between training and testing, what is the most likely effect on model accuracy?

A Accuracy will drop significantly due to label mismatch

B Accuracy will improve because of new label mapping

C Accuracy will randomly fluctuate without pattern

D Accuracy remains unchanged as encoding does not affect labels

💡 Hint

Think about how the model interprets numeric labels during prediction.

If label encoding changes, the model's predictions will not match the true labels, causing accuracy to drop.

🔧 Debug

advanced

Identify the error in label encoding usage

What error will the following code produce and why?

ML Python

from sklearn.preprocessing import LabelEncoder

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

A Raises a ValueError because labels contain strings

B Prints [0, 1, 2] as labels are encoded correctly

C Raises a NotFittedError because transform is called before fit

D Prints an empty list because labels are not fitted

💡 Hint

Check the order of method calls for LabelEncoder.

LabelEncoder requires fit() before transform(). Calling transform() first raises NotFittedError.

❓ Model Choice

expert

Choosing encoding method for ordinal categorical feature

You have a categorical feature representing education levels: ['High School', 'Bachelor', 'Master', 'PhD']. Which encoding method is best to preserve the order information for a machine learning model?

A Label encoding to assign increasing integers to levels

B One-hot encoding to create separate binary columns

C Random encoding to assign random numbers to categories

D Frequency encoding to replace categories with their counts

💡 Hint

Think about preserving the natural order of categories.

Label encoding preserves order by assigning increasing integers, which helps models understand ordinal relationships.