Feature engineering is the creation, transformation and selection of input variables so that a model can capture the patterns relevant to its prediction task. Questions ask about binning continuous values, log and polynomial transforms, interaction features and extracting date parts or lag values for time-series forecasting. Categorical encoding is tested through one-hot, ordinal, frequency and target encoding and their risks. Selection items compare filter methods such as correlation and chi-square, wrapper methods like recursive feature elimination, and embedded methods like Lasso. A recurring theme is target leakage, where a feature secretly contains information about the label that will not exist at prediction time.
Below are 30 practice questions from a pool of 210 Feature Engineering MCQs, one of 17 topics in AI / Machine Learning & Data Analytics. Each shows the correct answer with an explanation; when you are ready, take a timed quiz to test recall under exam conditions.
Each question below shows the correct answer with a full explanation. Use these to build conceptual understanding before attempting a timed quiz.
Q1. What is feature engineering?
Feature engineering is the process of creating, transforming, or selecting features from raw data to improve ML model performance.
Q2. Why might you create interaction features?
Interaction features capture the combined effect of two or more features that may not be apparent when considering them individually.
Q3. What does dropping a feature mean?
Dropping a feature means removing it from the dataset, typically because it is irrelevant, redundant, or harmful to model performance.
Q4. What is the purpose of polynomial features?
Polynomial features create new features by raising existing features to powers and computing their interactions, enabling linear models to capture non-linear patterns.
Q5. What is the Variance Inflation Factor (VIF)?
VIF measures how much a feature's variance is inflated due to correlation with other features. High VIF (>5-10) indicates problematic multicollinearity.
Q6. What is feature selection?
Feature selection identifies and keeps only the most relevant features, removing irrelevant or redundant ones to improve model performance.
Q7. What is a numerical feature?
Numerical features represent quantifiable measurements with continuous (e.g., height) or discrete (e.g., count) numeric values.
Q8. What is the purpose of log transformation on features?
Log transformation reduces right skewness, compresses large values, and can linearize multiplicative relationships between features and the target.
Q9. What is a datetime feature?
Datetime features contain date/time information from which useful features like day of week, month, hour, or is_weekend can be extracted.
Q10. What is a feature?
A feature is an individual measurable property or attribute of the data used as input for a machine learning model.
Q11. What is binning?
Binning (discretization) groups continuous values into discrete intervals or bins, like converting age into age groups.
Q12. What is target encoding?
Target encoding replaces each category with the mean of the target variable for observations in that category, useful for high-cardinality features.
Q13. What is a binary feature?
A binary feature takes only two values, typically 0 and 1, representing the presence or absence of a characteristic.
Q14. What is a lag feature in time series?
Lag features use values from previous time steps as input features, capturing temporal dependencies in time series data.
Q15. Feature extraction involves:
Feature extraction creates new features from raw data, such as extracting day of week from a date or word counts from text.
Q16. Why is feature scaling important for KNN?
KNN relies on distance metrics, so features with larger scales would dominate distance calculations, making scaling essential for fair comparison.
Q17. What is the Boruta algorithm for feature selection?
Boruta creates shadow (randomized) copies of all features, trains a Random Forest, and compares each feature's importance to the best shadow feature to determine relevance.
Q18. What is the difference between wrapper, filter, and embedded feature selection?
Filter methods use statistical measures independently of the model; wrapper methods evaluate feature subsets using a model; embedded methods perform selection during model training (e.g., Lasso).
Q19. What is feature hashing?
Feature hashing maps a large number of features to a fixed-size vector using hash functions, useful for high-dimensional sparse data like text.
Q20. What are rolling window features?
Rolling window features compute statistics (mean, std, etc.) over a sliding window of past observations, capturing recent trends in time series data.
Q21. What is domain knowledge in feature engineering?
Domain knowledge leverages understanding of the specific problem area to engineer features that capture meaningful patterns, often outperforming automated approaches.
Q22. What is the purpose of the chi-squared test in feature selection?
The chi-squared test measures the independence between categorical features and the target variable; features with high chi-squared values are more informative.
Q23. What is feature crossing in deep learning?
Feature crossing creates new features by combining existing ones (e.g., latitude x longitude), helping models learn interactions that may not be captured individually.
Q24. What is the difference between forward and backward feature selection?
Forward selection starts with no features and iteratively adds the most useful one. Backward elimination starts with all features and iteratively removes the least useful one.
Q25. What is mutual information in feature selection?
Mutual information measures how much knowing one variable reduces uncertainty about another, capturing both linear and non-linear dependencies.
Q26. How does LASSO perform feature selection?
LASSO's L1 penalty can shrink coefficients exactly to zero, effectively performing automatic feature selection by eliminating irrelevant features.
Q27. What is the curse of dimensionality's impact on feature engineering?
As dimensions increase, data becomes sparse and distance metrics become less meaningful, requiring exponentially more samples to maintain statistical significance.
Q28. What is permutation importance?
Permutation importance measures how much model performance decreases when a feature's values are randomly shuffled, indicating that feature's contribution to predictions.
Q29. What is automated feature engineering (e.g., Featuretools)?
Automated feature engineering tools like Featuretools use deep feature synthesis to automatically create features by applying transformation and aggregation primitives across related tables.
Q30. What is the concept of information leakage in feature engineering?
Information leakage occurs when features inadvertently contain information about the target that wouldn't exist in real-world prediction scenarios, leading to over-optimistic performance.
Practice MCQs with explanations
Practice MCQs with explanations
Practice MCQs with explanations
Practice MCQs with explanations
Practice MCQs with explanations
Practice MCQs with explanations
Take a timed quiz drawn from 210+ questions on this topic. No signup required — your progress saves in your browser.
Start Feature Engineering Quiz