3/26/2008

Introduction to Multiple Regression

• Our goal is to estimate a mean value of Y or predict an individual Y value based on a set of predictor variables.
• General form of multiple regression model
  • interaction terms
  • quadratic and higher order terms
• When there are p candidate predictor variables, how many models are possible?
• A model with too many predictor variables reduces the df associated with SSE and hence leads to a reduction in the precision of the inferences.
• Adjusted R-squared takes into account the number of predictor variables in the model.
• Don't forget to check your model assumptions, with a normal probability plot and residual plots.

Please read Chapter 14, including Chapter 14.pdf for class on Friday.