AML AML Feature Engineering & Data Preprocessing 1 — Questions and Answers
Question 1: What is the primary purpose of feature scaling in machine learning?
- To reduce the number of features
- To ensure all features contribute equally to model training (Correct answer)
- To remove outliers from the dataset
- To encode categorical variables
Correct answer: To ensure all features contribute equally to model training
Feature scaling normalizes feature ranges so that no single feature dominates model training due to its magnitude.
Question 2: Which technique handles missing values by replacing them with the average of the column?
- One-hot encoding
- Label encoding
- Mean imputation (Correct answer)
- Standardization
Correct answer: Mean imputation
Mean imputation replaces missing values with the column's average, preserving the dataset's overall distribution.
Question 3: What does one-hot encoding do to categorical variables?
- Assigns numeric labels to categories
- Creates binary columns for each category (Correct answer)
- Removes low-frequency categories
- Scales categories to a 0-1 range
Correct answer: Creates binary columns for each category
One-hot encoding converts each category into a separate binary column, avoiding the assumption of ordinal relationships.
Question 4: Which dimensionality reduction technique projects data onto directions of maximum variance?
- t-SNE
- LDA
- PCA (Correct answer)
- UMAP
Correct answer: PCA
PCA (Principal Component Analysis) finds orthogonal axes of maximum variance to compress feature dimensions.
Question 5: What is the primary benefit of feature selection in a machine learning pipeline?
- Increases model complexity
- Reduces overfitting and training time (Correct answer)
- Adds synthetic training data
- Improves data imputation accuracy
Correct answer: Reduces overfitting and training time
Feature selection removes irrelevant or redundant features, reducing overfitting risk and computational cost.
Question 6: Which outlier detection method uses the interquartile range (IQR) to flag extreme values?
- Z-score normalization
- Min-max scaling
- Winsorization
- Tukey's fence method (Correct answer)
Correct answer: Tukey's fence method
Tukey's fence method flags values beyond 1.5x IQR from Q1 or Q3 as outliers for removal or treatment.
AML AML Feature Engineering & Data Preprocessing 1
What is the primary purpose of feature scaling in machine learning?