1. . Feature engineering is:
(A) Encrypting features only
(B) The process of creating new input features from existing data to improve model performance
(C) Compressing datasets only
(D) Backup only
2. . Feature selection is:
(A) The process of choosing the most relevant features for model training
(B) Encrypting selected features
(C) Compressing features
(D) Backup only
3. . Why is feature engineering important?
(A) It helps improve model accuracy, interpretability, and efficiency
(B) Encrypts features only
(C) Compresses datasets
(D) Backup only
4. . Common feature engineering techniques include:
(A) Compressing features only
(B) Encrypting variables only
(C) Encoding categorical variables, scaling, creating interaction terms, and aggregations
(D) Backup only
5. . One-hot encoding is used to:
(A) Convert categorical variables into binary vectors
(B) Encrypt categories
(C) Compress data
(D) Backup only
6. . Label encoding is:
(A) Compressing labels
(B) Encrypting labels
(C) Converting categorical variables into numeric labels
(D) Backup only
7. . Feature scaling is important because:
(A) Encrypting scaling
(B) Many algorithms like SVM, KNN, and gradient descent are sensitive to feature magnitude
(C) Compressing features
(D) Backup only
8. . Standardization scales features by:
(A) Subtracting mean and dividing by standard deviation
(B) Encrypting data
(C) Compressing data
(D) Backup only
9. . Normalization scales features by:
(A) Compressing values
(B) Encrypting values
(C) Bringing values to a fixed range, often [0,1]
(D) Backup only
10. . Feature interaction involves:
(A) Creating new features by combining two or more existing features
(B) Encrypting features
(C) Compressing features
(D) Backup only
11. . Principal Component Analysis (PCA) is used for:
(A) Backup only
(B) Encrypting components
(C) Compressing features
(D) Dimensionality reduction by transforming features into uncorrelated components
12. . Recursive Feature Elimination (RFE) is:
(A) Encrypting features
(B) A method to select important features by recursively removing less important ones
(C) Compressing features
(D) Backup only
13. . Mutual information in feature selection measures:
(A) The dependency between features and the target variable
(B) Encrypting dependency
(C) Compressing data
(D) Backup only
14. . Correlation-based feature selection aims to:
(A) Backup only
(B) Encrypt correlations
(C) Compress features
(D) Remove redundant features that are highly correlated with others
15. . L1 regularization (Lasso) can be used for:
(A) Compressing coefficients
(B) Encrypting coefficients
(C) Feature selection by shrinking less important feature coefficients to zero
(D) Backup only
16. . Feature importance scores from tree-based models help to:
(A) Compress features
(B) Encrypt features
(C) Identify the most influential features for predictions
(D) Backup only
17. . Removing irrelevant or noisy features helps to:
(A) Encrypt features
(B) Improve model performance and reduce overfitting
(C) Compress datasets
(D) Backup only
18. . Dimensionality reduction helps to:
(A) Backup only
(B) Encrypt features
(C) Compress features
(D) Reduce computational cost while retaining important information
19. . Interaction terms are useful when:
(A) Encrypting interactions
(B) Relationships between features affect the target variable
(C) Compressing interactions
(D) Backup only
20. . The main purpose of feature engineering and selection is to:
(A) Encrypt features only
(B) Prepare high-quality, relevant features to improve model accuracy, efficiency, and interpretability
(C) Compress datasets only
(D) Backup only