Multiple regression is the technique of fitting or predicting a response variable from a linear combination of several other variables. The fitting principle is least squares, the same as with simple linear regression.
Many regression concepts were introduced in previous chapters, so this chapter concentrates on showing some new concepts not encountered in simple regression: the point-by-point picture of a hypothesis test with the leverage plot, collinearity (the situation in which one regressor variable is closely related to another), and the case of exact linear dependencies.