Topic 13

Question 1

Multiple Regression

Answer 1

Linear regression with more than 2 variables (eg. book- weight, volume, paper vs hardback)

Question 2

Factor

Answer 2

Categorical variable predictor

Question 3

Reference Level

Answer 3

Levels are the different cateogorical values the factor can take, reference level coded as 0 other levels coded as 1 to compare

Question 4

Collinear Explanatory Variables

Answer 4

Two predictor variables that are correlated, complicates the model estimation

Question 5

Adjusted R^2

Answer 5

Accounts for the number of explanatory variables, applying a penalty for the number of preditors

Question 6

Feature Seletction: Backwards Elimination- Adjusted R^2 Approach

Answer 6

• Start with the full model
• Drop one variable at a time and record adjusted R2 of each smaller model
• Pick the model with the highest increase in adjusted R2
• Repeat until none of the smaller models yield an increase in adjusted R2

Question 7

Feature Selection: Backwards Elimination- p-value Approach

Answer 7

• Start with the full model
• Drop the variable with the highest p-value and refit a smaller model
• Repeat until all variables left in the model are significant

Question 8

Feature Selection: AIC Approach

Answer 8

*
Similar to the adjusted R2 approach: penalises for adding more variables to the model.
*
AIC quantifies the amount of information loss due to simplification of the model (less information loss is better)