Linear Regression by Least Squares

Abstract

We usually study more than one variable at a time. When the variables are continuous, and one is clearly a response variable and the others are predictor variables, we usually plot the variables and then attempt to fit a model to the plotted points. With one continuous predictor, the first model we attempt is a straight line; with two or more continuous predictors, we attempt a plane. We plot the model, the residuals from the model, and various diagnostics of the quality of the fit.