Lasso Regression Overview Answers 1. What is a potential disadvantage of using Lasso regression? A. It may yield suboptimal coefficient estimates in some situations. B. It offers more stable model predictions compared to Ridge Regression. C. It provides optimal coefficient estimates in all scenarios. D. It is designed specifically for low-dimensional data analysis. Answer: It may yield suboptimal coefficient estimates in some situations. (A) Lasso regression can provide suboptimal coefficient estimates in specific scenarios, making it less reliable in certain contexts. 2. In which area is Lasso regression commonly applied? A. Real estate valuation B. Genomics for identifying important genes C. Physical therapy assessment D. Environmental impact studies Answer: Genomics for identifying important genes (B) Lasso regression is used in genomics to help identify significant genes involved in complex biological processes. 3. What model evaluation metric is commonly used to assess the performance of a Lasso regression model? A. Correlation Coefficient B. Mean Absolute Deviation C. Mean Squared Error D. Variance Inflation Factor Answer: Mean Squared Error (C) Mean Squared Error (MSE) is a key metric used to evaluate the accuracy and fit of Lasso regression models. 4. What behavior might Lasso regression exhibit when dealing with highly correlated predictors? A. It may yield very precise predictions. B. It enhances the selection of redundant features. C. It consistently selects all correlated features. D. It can behave unpredictably with feature selection. Answer: It can behave unpredictably with feature selection. (D) Lasso regression can behave unpredictably and select correlated features inconsistently when facing highly correlated predictors. 5. Which software packages provide implementations of Lasso regression? A. SAS and SPSS B. scikit-learn and R C. Photoshop and Illustrator D. MATLAB and Excel Answer: scikit-learn and R (B) Packages like scikit-learn and R include Lasso regression, making it accessible for users working with statistical models. 6. What is the primary purpose of Lasso regression? A. To perform variable selection and regularization B. To eliminate the need for cross-validation C. To minimize the coefficient counts D. To increase model complexity Answer: To perform variable selection and regularization (A) Lasso regression aims to minimize error while also performing variable selection and regularization. 7. Which regularization method does Lasso use? A. L2 regularization B. No regularization C. Elastic net regularization D. L1 regularization Answer: L1 regularization (D) Lasso uses L1 regularization, which penalizes the sum of the absolute values of the regression coefficients. 8. What effect does increasing the tuning parameter λ have in Lasso regression? A. Reduces bias in the model B. Increases variance in the model C. Leads to more coefficients being shrunk to zero D. Decreases model interpretability Answer: Leads to more coefficients being shrunk to zero (C) Larger values of λ lead to more variables being set to zero, increasing the shrinkage effect. 9. Why is model generalization improved in Lasso regression? A. It uses more predictors than observations B. It reduces bias completely C. It only focuses on training data D. It introduces bias to reduce variance Answer: It introduces bias to reduce variance (D) Lasso introduces bias into the model, which reduces variance and improves generalization. 10. What is one key characteristic that distinguishes Lasso regression from Ridge regression? A. Ability to eliminate coefficients B. Use of L2 regularization C. Dependency on a migration algorithm D. Focus on training data only Answer: Ability to eliminate coefficients (A) Lasso regression can effectively eliminate coefficients by shrinking some to exactly zero, while Ridge regression does not. 11. In the context of selecting the optimal tuning parameter λ, what method is commonly used? A. Training with a fixed λ value B. Randomized variable selection C. Cross-validation D. Using only training data Answer: Cross-validation (C) Cross-validation is a common method for selecting the optimal λ value by evaluating model performance. 12. What advantage does Lasso provide when dealing with high-dimensional data? A. Reduces feature selection to a single predictor B. Simplifies the model and enhances performance C. Eliminates overfitting by using all available variables D. Increases the number of predictors Answer: Simplifies the model and enhances performance (B) Lasso effectively simplifies the model, making it easier to interpret while improving performance in highdimensional data. 13. What happens to the model complexity when Lasso regression is applied? A. Complexity is reduced by selecting relevant variables B. Complexity increases due to many predictors C. Complexity decreases regardless of λ D. Complexity remains unchanged Answer: Complexity is reduced by selecting relevant variables (A) Lasso simplifies the model by selecting a subset of relevant variables, thereby reducing model complexity.