1. (5 points) What is the null and alternative Hypothesis of the F-test? …of a t-test? Explain how each one can be used in the analysis of your regression model.
2. (5 points) What are the four assumptions about residuals in the regression model? Why are these assumptions made? How can you verify your assumptions? How can you correct your model if the assumptions are not verified?
3. (5 points) How can you judge the quality of a model? What metrics can you use to compare models?
4. (5 points) Given a model that predicts y given x1 and x2 write the a) first order model, b) interaction model and c) complete second order model. Which is better, under which circumstances?
5. (5 points) In the model below, what is Beta-0, Beta-1, Beta-2? What is the regression line? Why was this line chosen? What is the SSE? Can you be certain that x1 and x2 should be in the model? What is R2? What does that mean? What is MSE? What does that mean? RMSE? What does that mean?
6. (5 points) How can you validate your model? Give two distinctly different methods?
7. (5 points) Explain as if to a nonprofessional why adjusted-R2 might be better than R2.
8. (5 points) Define "parsimonious." Explain its relevance to building regression models.
9. (5 points) Explain how to incorporate categorical features into your model? Be specific.
10. (5 points) Compare and contrast the benefits and drawbacks of forward stepwise regression, backward stepwise regression, and all-possible regression.
