One-hot encoding in ML Python - Practice Problems & Coding Challenges

Challenge - 5 Problems

❓ Predict Output

intermediate

What is the output of this one-hot encoding code?

Consider this Python code using pandas to one-hot encode a categorical column. What is the printed output?

ML Python

import pandas as pd

data = {'color': ['red', 'blue', 'green', 'blue']}
df = pd.DataFrame(data)
dummies = pd.get_dummies(df['color'])
print(dummies)

A

   red  blue  green
0    1     0      0
1    0     1      0
2    0     0      1
3    0     1      0

B

   blue  green  red
0     0      0    1
1     1      0    0
2     0      1    0
3     1      0    0

C

   green  red  blue
0      0    1     0
1      0    0     1
2      1    0     0
3      0    0     1

D

   red  green  blue
0    1      0     0
1    0      0     1
2    0      1     0
3    0      0     1

💡 Hint

Look at the order of columns created by pd.get_dummies and how the rows correspond to the original data.

pd.get_dummies creates columns in sorted order of unique values. Here, 'blue', 'green', 'red' are sorted alphabetically. Each row has 1 in the column matching its color and 0 elsewhere.

🧠 Conceptual

intermediate

Why use one-hot encoding for categorical data?

Which of the following is the main reason to use one-hot encoding on categorical features before training a machine learning model?

A To convert categories into numbers so the model can process them without assuming order.

B To reduce the number of features and simplify the model.

C To normalize the data between 0 and 1 for better training.

D To combine multiple categories into a single numeric value.

💡 Hint

Think about how models interpret numeric inputs and what happens if categories are just assigned numbers.

One-hot encoding converts categories into separate binary columns so models do not assume any order or priority among categories.

❓ Hyperparameter

advanced

Choosing the right approach for high-cardinality categorical features

You have a categorical feature with 10,000 unique values. Which one-hot encoding approach is best to avoid excessive memory use?

A Use standard one-hot encoding creating 10,000 binary columns.

B Use label encoding assigning integers to categories.

C Use one-hot encoding but drop one category to reduce columns by one.

D Use target encoding to replace categories with average target values.

💡 Hint

Think about memory and model performance when many categories exist.

Standard one-hot encoding creates too many columns, causing memory and performance issues. Target encoding reduces dimensionality by using target statistics.

❓ Metrics

advanced

Effect of one-hot encoding on model accuracy

You train two models on the same dataset: Model A uses label encoding for a categorical feature, Model B uses one-hot encoding. Model B shows better accuracy. Why?

A Label encoding always causes models to overfit, reducing accuracy.

B One-hot encoding reduces the number of features, making training faster and more accurate.

C One-hot encoding prevents the model from assuming an order in categories, improving accuracy.

D Label encoding converts categories to strings, which models cannot process.

💡 Hint

Consider how models interpret numeric values from label encoding.

Label encoding assigns arbitrary numeric values that may imply order, misleading the model. One-hot encoding avoids this by treating categories independently.

🔧 Debug

expert

Why does this one-hot encoding code raise an error?

What error does this code raise and why? import numpy as np categories = ['cat', 'dog', 'bird'] values = ['cat', 'dog', 'fish'] one_hot = np.zeros((len(values), len(categories))) for i, val in enumerate(values): idx = categories.index(val) one_hot[i, idx] = 1 print(one_hot)

A ValueError: 'fish' is not in list because 'fish' is not in categories.

B IndexError: index out of range because 'fish' index is too large.

C TypeError: unsupported operand type(s) because of wrong data types.

D No error, prints a one-hot encoded numpy array.

💡 Hint

Check if all values exist in the categories list before indexing.

The code tries to find the index of 'fish' in categories, but 'fish' is not present, causing a ValueError.