Regression I

Lecture 2

Simple linear example, fixed x values, imagine normal distribution of y's at each x
    Expected value of y at each x
    Distribution of epsilons
    Forms of model equations (p 8-9), population equations vs sample equations
    Expected values (this means means)

Exploratory vs Confirmatory Data Analysis (p. 14)

Variance--average squared deviation from the mean.  Pop vs Sample. Why n-1?

Proof of alternative formula for variance (sum of squares) just for some practice.

How do we model the mean in regression?   Insert into variance formula.  Assume constant variance!

Methods of fitting a line to data
   Eyeball
   Minimize distances (Mean Absolute Deviations)
   Minimize sum of squared errors
      What is y-hat? (equation)