fit.model: Fitting the model
In sumitrahman/perfman: Probability Estimates From Ordered Logistic Regression

Description Usage Arguments Details Value Examples

fit.model fits the ordered logistic model and sets up some useful variables for the other perfman functions.

1	fit.model(dataset)

dataset

A data frame whose first column is an ordered numeric factor containing the positive integers and whose subsequent columns are the explanatory variables for modelling. For analysis in this package all the explanatory variables must be factors.

The function fits the model using MASS::polr. This is saved to fit.

Then the vector cuts is created which are the cuts from the model (these divide the cumulative density into the areas set for the ordered outcomes), including -Inf for the zeroth cut and Inf for the end point. This makes it convenient for another perfman function to use.

Next two sets of tables are created which create the marginal distributions from the dataset that is being analysed. counts.raw gives you the actual counts, while proportions.raw gives you the proportions.

Finally a vector of variables is created, again for the convenience of other functions in perfman. This vector simply contains the names of the explanatory variables from the model.

An object of class polr. See documentation for the polr function in the MASS package for more details.