Regression Overview
see also

Regression analysis is used to investigate and model the relationship between a response variable and one or more predictors. Minitab provides the following regression procedures:

· Least squares: Use when your response variable is continuous.

· Nonlinear regression: Use when you cannot adequately model the relationship with linear parameters.

· Orthogonal regression: Use when the response and predictor both contain measurement error.

· Partial least squares regression: Use when your predictors are highly correlated or outnumber your observations.

· Logistic regression: Use when your response variable is categorical.

· Poisson regression: Use when your response variable counts occurrences.

Both least squares and logistic regression methods estimate parameters in the model so that the fit of the model is optimized. Least squares methods minimize the sum of squared errors to obtain parameter estimates, whereas Minitab's logistic regression obtains maximum likelihood estimates of the parameters. See Generalized Linear Models Overview for more information. Partial least squares (PLS) extracts linear combinations of the predictors to minimize prediction error. See Partial Least Squares Overview for more information.

Use the table below to select a procedure:

Use...	To...	Response type	Estimation method
Regression	perform simple, multiple, polynomial, and stepwise least squares regression with continuous and categorical predictors	continuous	least squares
Best Subsets	identify subsets of the predictors based on the maximum R $image\SQUARED.gif$ criterion	continuous	least squares
Fitted Line Plot	perform linear and polynomial regression with a single predictor and plot a regression line through the data	continuous	least squares
Nonlinear Regression	perform simple or multiple regression using the nonlinear function of your choice	continuous	least squares
Stability Study	analyze the stability of a product over time and determine the product's shelf life	continuous	least squares, Restricted Maximum Likelihood (REML)
Orthogonal Regression	perform orthogonal regression with one response and one predictor	continuous	orthogonal
PLS	perform regression with ill-conditioned data	continuous	biased, non-least squares
Binary Logistic	perform logistic regression on a response with only two possible values, such as presence or absence	categorical	maximum likelihood
Ordinal Logistic	perform logistic regression on a response with three or more possible values that have a natural order, such as none, mild, or severe	categorical	maximum likelihood
Nominal Logistic	perform logistic regression on a response with three or more possible values that have no natural order, such as sweet, salty, or sour	categorical	maximum likelihood
Poisson Regression	perform Poisson regression on a response that counts occurrences	discrete	maximum likelihood

Regression Overview see also

Regression Overview
see also