ABSTRACT
Starting with investigating the relationship between the explanatory and outcome variables, this chapter first reviews parametric methods that can be applied to quantify the effect of a set of explanatory variables of a particular outcome. It introduces several nonparametric regression methods in which there is no parametric form assumed for the relationship between the outcome and the explanatory variables. Regression analysis traces the average value of a response variable (also known as a dependent variable) as a function of one or several explanatory variables (also known as covariates, predictors, or independent variables). Graphical analysis of residuals (original or scaled) is a very effective way to investigate the adequacy of the fit. These include the normal probability plot of residuals, the plot of residuals against the fitted values, the plot of residuals against each regressor variable, and the plot of residuals in time series if time series data were collected.
