Ordinal Logistic Regression - Options
main topic     see also  

Stat > Regression > Ordinal Logistic Regression > Options

Provides options that allow you to choose a link function, the reference level and event, and specify starting values.

Dialog box items

Link Functions Minitab provides three link functions, allowing you to fit a broad class of binary response models. The link function is the inverse of a distribution function. See Link Functions for more information.

Logit: Choose to use the logit link function. This is the default.

Normit/Probit: Choose to use the normit link function.

Gompit/Complementary log-log: Choose to use the gompit link function (also called the complementary log-log function). The gompit function is the inverse of the Gompertz distribution function.

Reference Options

Order of the response values: Enter the order of the response values from lowest to highest if you want to change the reference event. See Interpreting the parameter estimates.

Reference factor level (enter factor followed by level): Specify the reference factor level by entering the factor column followed by the reference level. (Text and date/time levels must be enclosed in quotes.) See Interpreting the parameter estimates.

Algorithm Options

Starting estimates for algorithm: Specify the column containing the initial values for model parameters. The column containing the initial values must have one row for each estimated coefficient in the model. The starting values for all the constants appear first, followed by starting values the predictors in the model. There must be one row for each degree of freedom.

Estimates for validation model: Specify the column containing the estimated model parameters. Minitab will then fit the validation model.

Maximum number of iterations: Enter the maximum number of iterations you want Minitab to perform to reach convergence. The default value is 20. Minitab obtains maximum likelihood estimates through an iterative process. If the maximum number of iterations is reached before convergence, the command terminates.