Data Science Data Science Supervised Learning Models Questions and Answers 2 — Questions and Answers
Question 1: Which supervised learning algorithm constructs a series of if-then rules by recursively partitioning the feature space?
- Decision Tree (Correct answer)
- K-Means Clustering
- Principal Component Analysis
- DBSCAN
Correct answer: Decision Tree
Decision trees split data recursively using feature thresholds to create interpretable if-then decision rules.
Question 2: What is the primary purpose of regularization in supervised learning models such as Ridge and Lasso regression?
- To prevent overfitting by penalizing large coefficients (Correct answer)
- To increase the number of features in the model
- To convert categorical variables into numerical ones
- To speed up the training process
Correct answer: To prevent overfitting by penalizing large coefficients
Regularization adds a penalty term to the loss function that discourages overly complex models with large coefficient values.
Question 3: In a Random Forest classifier, what technique is used to reduce variance and improve generalization?
- Bagging with random feature subsets (Correct answer)
- Gradient boosting with sequential trees
- Forward feature selection
- Backward elimination of correlated features
Correct answer: Bagging with random feature subsets
Random Forest combines bagging (bootstrap aggregating) with random feature selection at each split to decorrelate trees and reduce variance.
Question 4: Which metric is most appropriate for evaluating a supervised classification model when the dataset has a severe class imbalance?
- Area Under the Precision-Recall Curve (AUPRC) (Correct answer)
- Accuracy
- Mean Squared Error
- R-squared
Correct answer: Area Under the Precision-Recall Curve (AUPRC)
AUPRC focuses on the performance for the minority class and is more informative than accuracy when classes are heavily imbalanced.
Question 5: What does the kernel trick enable Support Vector Machines to do?
- Map data into a higher-dimensional space to find nonlinear decision boundaries (Correct answer)
- Reduce the number of support vectors needed
- Automatically select the best regularization parameter
- Perform unsupervised clustering on the data
Correct answer: Map data into a higher-dimensional space to find nonlinear decision boundaries
The kernel trick implicitly maps input features into a higher-dimensional space where a linear separator can capture nonlinear relationships.
Question 6: In gradient boosting, what does each successive tree in the ensemble attempt to model?
- The residual errors of the previous ensemble (Correct answer)
- A random subset of the original target values
- The inverse of the previous tree's predictions
- The feature importances from the prior iteration
Correct answer: The residual errors of the previous ensemble
Each new tree in gradient boosting is trained to predict the residual errors left by the combined predictions of all preceding trees.
Data Science Data Science Supervised Learning Models Questions and Answers 2
Which supervised learning algorithm constructs a series of if-then rules by recursively partitioning the feature space?