Label encoding in ML Python - ML Experiment: Train & Evaluate

Experiment - Label encoding

Problem:You have a dataset with categorical labels like colors: ['red', 'green', 'blue', 'green', 'red']. You want to convert these text labels into numbers so a machine learning model can understand them.

Current Metrics:N/A - The problem is about data preprocessing, not model accuracy yet.

Issue:The model cannot work directly with text labels. Without encoding, it will throw errors or perform poorly.

Your Task

Convert the categorical labels into numeric form using label encoding so they can be used in a machine learning model.

Use only label encoding (no one-hot encoding).

Use sklearn's LabelEncoder.

Keep the order of labels as in the original list.

Hint 1

Hint 2

Hint 3

Solution

ML Python

from sklearn.preprocessing import LabelEncoder

labels = ['red', 'green', 'blue', 'green', 'red']

encoder = LabelEncoder()
encoded_labels = encoder.fit_transform(labels)

print('Original labels:', labels)
print('Encoded labels:', encoded_labels.tolist())

Imported LabelEncoder from sklearn.preprocessing.

Created a LabelEncoder instance.

Fitted the encoder on the original labels.

Transformed the labels into numeric form.

Results Interpretation

Before encoding: ['red', 'green', 'blue', 'green', 'red']

After encoding: [2, 1, 0, 1, 2]

Label encoding converts text labels into numbers so models can process them. Each unique label gets a unique number.

Bonus Experiment

Try using one-hot encoding on the same labels and compare the results with label encoding.

💡 Hint

Use sklearn's OneHotEncoder and observe how the output changes from single numbers to arrays representing each category.

Practice

1. What is the main purpose of label encoding in machine learning?

easy

A. Convert categorical labels into numbers for model input

B. Normalize numerical data to a 0-1 range

C. Split data into training and testing sets

D. Reduce the number of features in the dataset

Solution

1. Step 1: Understand label encoding function

   Label encoding changes categories like 'red', 'blue' into numbers like 0, 1 so models can process them.

2. Step 2: Compare with other options

   Normalization scales numbers, splitting divides data, and feature reduction removes features, none are label encoding.

3. Final Answer:

   Convert categorical labels into numbers for model input -> Option A

4. Quick Check:

   Label encoding = Convert categories to numbers [OK]

Hint: Label encoding turns words into numbers for models [OK]

Common Mistakes:

• Confusing label encoding with normalization
• Thinking label encoding splits data
• Mixing label encoding with feature selection

2. Which of the following is the correct way to import and use LabelEncoder from scikit-learn in Python?

easy

A. from sklearn import LabelEncoder
encoded = LabelEncoder.fit(['cat', 'dog', 'cat'])

B. import LabelEncoder from sklearn
encoded = LabelEncoder(['cat', 'dog', 'cat'])

C. from sklearn.preprocessing import LabelEncoder
encoder = LabelEncoder()
encoded = encoder.fit_transform(['cat', 'dog', 'cat'])

D. from sklearn.preprocessing import LabelEncoder
encoded = LabelEncoder.transform(['cat', 'dog', 'cat'])

Solution

1. Step 1: Check import syntax

   The correct import is from sklearn.preprocessing import LabelEncoder.

2. Step 2: Check usage of fit_transform

   LabelEncoder requires creating an instance, then calling fit_transform on data.

3. Final Answer:

   from sklearn.preprocessing import LabelEncoder encoder = LabelEncoder() encoded = encoder.fit_transform(['cat', 'dog', 'cat']) -> Option C

4. Quick Check:

   Correct import and fit_transform usage [OK]

Hint: Import from sklearn.preprocessing and use fit_transform() [OK]

Common Mistakes:

• Wrong import path for LabelEncoder
• Calling transform without fit
• Using LabelEncoder as a function directly

3. What will be the output of this Python code using LabelEncoder?

from sklearn.preprocessing import LabelEncoder
encoder = LabelEncoder()
labels = ['apple', 'banana', 'apple', 'orange']
encoded_labels = encoder.fit_transform(labels)
print(list(encoded_labels))

medium

A. [0, 1, 0, 2]

B. [1, 2, 1, 3]

C. [0, 0, 1, 2]

D. [1, 0, 1, 2]

Solution

1. Step 1: Identify unique labels and their order

   Unique labels sorted alphabetically are ['apple', 'banana', 'orange'].

2. Step 2: Assign numbers based on alphabetical order

   'apple' = 0, 'banana' = 1, 'orange' = 2, so encoded list is [0,1,0,2].

3. Final Answer:

   [0, 1, 0, 2] -> Option A

4. Quick Check:

   Alphabetical order encoding = [0,1,0,2] [OK]

Hint: LabelEncoder assigns numbers alphabetically [OK]

Common Mistakes:

• Assuming order of appearance instead of alphabetical
• Mixing up label indices
• Forgetting to convert to list before printing

4. You run this code but get an error:

from sklearn.preprocessing import LabelEncoder
encoder = LabelEncoder()
labels = ['red', 'blue', 'green']
encoded = encoder.transform(labels)
print(encoded)

What is the problem?

medium

A. transform() only works on numbers, not strings

B. LabelEncoder cannot encode color names

C. You should import LabelEncoder from sklearn.preprocessing.label

D. You must call fit or fit_transform before transform

Solution

1. Step 1: Understand LabelEncoder usage

   LabelEncoder requires fitting on data before transforming new data.

2. Step 2: Identify missing fit step

   The code calls transform without fit or fit_transform, causing error.

3. Final Answer:

   You must call fit or fit_transform before transform -> Option D

4. Quick Check:

   fit before transform = required [OK]

Hint: Always fit before transform with LabelEncoder [OK]

Common Mistakes:

• Calling transform without fitting first
• Wrong import path
• Thinking transform works on raw strings directly

5. You have a dataset with a categorical feature 'Fruit' containing ['apple', 'banana', 'apple', 'banana', 'orange', 'banana']. You want to encode it for a model that treats numbers as ordered values. Which approach is best?

hard

A. Use LabelEncoder to assign numbers (0,1,2) to fruits

B. Manually assign numbers based on fruit sweetness order

C. Use OneHotEncoder to create separate binary columns for each fruit

D. Leave the feature as text because encoding is not needed

Solution

1. Step 1: Understand model needs for ordered values

   The model treats numbers as ordered, so encoding must reflect meaningful order.

2. Step 2: Evaluate encoding options

   LabelEncoder assigns arbitrary numbers alphabetically, OneHotEncoder creates separate columns without order, manual assignment can reflect sweetness order.

3. Step 3: Choose best approach

   Manual assignment based on domain knowledge preserves order, fitting model assumptions.

4. Final Answer:

   Manually assign numbers based on fruit sweetness order -> Option B

5. Quick Check:

   Ordered encoding needs meaningful number assignment [OK]

Hint: Assign numbers reflecting real order for ordered models [OK]

Common Mistakes:

• Using LabelEncoder blindly for ordered data
• Confusing one-hot with ordered encoding
• Ignoring model assumptions about number meaning