OLS

Question 1

Best Linear Predictor of OLS

Answer 1

An estimand that represents the linear function of X that best predicts Y. Also shown as (y= a +bX). It minimizes the sum of the squared prediction errors (SSR)

Question 2

Does linearity mean a straight line?

Answer 2

No, but it does mean that of the beta coefficients are to the power of 1 and are not squared

Question 3

How to use the ordinary least squares estimator?

Answer 3

First, you sample units from the target population and collect data on both X and Y for each unit. Then, you find the a and b that minimize the SSR. Lastly, you compute the SSR for each combination of a and b, and then you pick the combination that yields the smallest r.

Question 4

Correlation

Answer 4

The extent to which two features of the world tend to occur together. This must be variation on the two features. And the best linear predictor is a way to capture correlation.