Regression Part-5 Classical Assumptions

Question 1

What is the linearity assumption?

Answer 1

It has to be linear in parameters (coefficients) but not linear in variables

Question 2

How can you detect if the functional form is incorrect?

Answer 2

Use residual plots

Question 3

What are the 10 assumptions of CLRM?

Answer 3

1. Linearity in parameters
2. X independent of error term cov(xi, ui) = 0
3. zero mean value of ui
4. Homoskedasticity (var(ui) = sigma squared)
5. No autocorrelation between the ui
6. Number of observations > number of parameters
7. X variable and values must be different; no outliers
8. No multicollinearity
9. No specification bias
10. ui is normally distributed

Question 4

What is BLUE

Answer 4

Linear, unbiased (expected value fo predicted estimator is same as true value) and effiecient estimator (min variance )