ABSTRACT
Μultivariable regression is a very popular statistical technique used to express a random variable, known as the dependent variable, as a linear combination of a number of random variables, known as the independent variables. The simple linear regression model with one independent random variable is first described, followed by the general case of multivariable regression. A discussion on residual analysis and other measures of goodness of fit used to determine the accuracy of a regression model, is then given. Subsequently, polynomial regression, confidence and prediction intervals, and the ridge, lasso, and elastic net regression techniques, are presented. A set of exercises and a regression project is given at the end of the Chapter.
