CDSP 6 Quiz

1. What is the purpose of training a regression model?

A: To classify data into different categories
B: To predict the relationship between dependent and independent variables
C: To cluster similar data points together
D: To reduce the dimensionality of data

2. In a linear equation, what does the 'm' represent?

A: The dependent variable
B: The slope of the line
C: The intercept of the line
D: The independent variable

3. What is a key limitation of linear regression?

A: It cannot be used with more than one predictor variable
B: It does not work well with non-linear data
C: It requires large amounts of data
D: It can only be used for classification problems

4. What does the ŷ symbol represent in a regression model?

A: The independent variable
B: The actual observed value
C: The estimated or predicted value
D: The model parameter

5. What method is commonly used to solve for parameters in a linear regression model?

A: Decision trees
B: The normal equation
C: Principal Component Analysis (PCA)
D: Gradient boosting

6. What is the Moore-Penrose inverse used for in linear regression?

A: Handling non-square matrices
B: Solving the normal equation
C: Calculating gradient descent
D: Scaling the features

7. What is one advantage of regression trees over linear regression models?

A: Regression trees are less prone to overfitting
B: Regression trees can handle non-linear relationships
C: Regression trees require fewer data points
D: Regression trees are easier to interpret

8. What error metric is commonly used in regression trees?

A: Gini index
B: Mean squared error (MSE)
C: Precision
D: F1 score

9. What is the primary goal of ensemble learning in regression models?

A: To reduce the training time
B: To combine multiple models for more accurate predictions
C: To eliminate irrelevant features
D: To handle missing data

10. What does the ARIMA model stand for?

A: Autoregressive Integrated Moving Average
B: Autocorrelation Integrated Maximum Algorithm
C: Automated Regression In Machine Learning
D: Average Regression Interpolated Model Accuracy

11. In ARIMA models, what does the 'p' parameter represent?

A: The number of times the data is differenced
B: The number of lag observations included in the model
C: The number of moving average terms
D: The length of the time window

12. What is the purpose of applying regularization to a regression model?

A: To increase model accuracy
B: To penalize model complexity and reduce overfitting
C: To increase the learning rate
D: To scale the data

13. Which regularization technique minimizes the sum of squared coefficients?

A: Lasso
B: Ridge
C: Elastic Net
D: Gradient Boosting

14. What is the difference between Lasso and Ridge regularization?

A: Lasso eliminates irrelevant features, while Ridge reduces the influence of all features
B: Ridge eliminates features, while Lasso reduces the influence of all features
C: Lasso increases model complexity, while Ridge simplifies it
D: Lasso uses squared error, while Ridge uses absolute error

15. What is the main purpose of gradient descent in regression models?

A: To reduce the model’s size
B: To optimize the model by minimizing the cost function
C: To regularize the model
D: To adjust the learning rate

16. What is a key advantage of using mini-batch gradient descent over batch gradient descent?

A: Mini-batch requires fewer parameters
B: Mini-batch updates model weights faster
C: Mini-batch uses the entire dataset for each update
D: Mini-batch decreases the model complexity

17. What is the mean squared error (MSE) used for in regression?

A: To measure the variance of the dataset
B: To calculate the difference between predicted and actual values
C: To determine the accuracy of classification models
D: To identify feature importance

18. What is the primary benefit of using root mean squared error (RMSE)?

A: It eliminates outliers from the dataset
B: It makes the error metric easier to interpret
C: It increases the model’s accuracy
D: It reduces the dimensionality of the dataset

19. What does the R² value represent in regression models?

A: The percentage of total variation in the dependent variable explained by the model
B: The sum of squared errors between predicted and actual values
C: The total number of features used in the model
D: The likelihood of overfitting

20. What is one reason why a high R² value might not always indicate a good model?

A: It means the model has too few features
B: It may indicate the model is overfitting to the training data
C: It only measures accuracy, not precision
D: It suggests the model is underfitting the data