Provide brief reponses to the following prompts.

Sketch a representative $QQ$ plot for right skewed data when the the$-$

oretical quantiles correspond to a normal distribution.

Explain what we mean by "Data Splitting" in the context of model val$-$

idation.

List two criteria that can be used with automated variable selection pro$-$

cedures like Forward Selection and Backward Elimination $($their names

suffice $-$ you do not need to provide the formulas$).$ For each criterion,

specify if a high or low value is desired.

Discuss why Forward Selection and Backward Elimination would not

work when $R^2$ is used as the performance criterion.

Hint: For $(4)$ consider the difference between $R_a^2$ and $R^2.$ What motivated

us to introduce $R_a^2$ in class?