Linear regression

Question 1

What is a model?

Answer 1

Result of probability distributions, can be used to predict most values a sample will take on.

Question 2

What do models predict?

Answer 2

How a change in the independent variable affects a dependent variables

Question 3

What is the model error?

Answer 3

Difference between model and outlying observed value

Question 4

What is the coefficient of determination?

Answer 4

R^2 = 1 - SSerror/SStotal

Used for model validation.

Compares the variance of the error of the model against the variance of the sample itself.

Question 5

How is the result of the coefficient of determination interpreted?

Answer 5

Lower the ratio, the less variance is explained by the model.

If ratio == 1 then 0 variance is explained by the model, making it worthless.

Question 6

How do we define SSerror?

Answer 6

Sum of model error (variance from model mean) against each value

Question 7

How do we define SStotal

Answer 7

Sum of error in the sample against it’s mean