yates_setup Method for adding new models to the yates function. 
 Description
This is a method which is called by the yates function, in order to setup the code to handle a particular model type. Methods for glm, coxph, and default are part of the survival package. 
Usage
yates_setup(fit, ...)
Arguments
| fit | a fitted model object | 
| ... | optional arguments for some methods | 
Details
If the predicted value should be the linear predictor, the function should return NULL. The yates routine has particularly efficient code for this case. Otherwise it should return a prediction function or a list of two elements containing the prediction function and a summary function. The prediction function will be passed the linear predictor as a single argument and should return a vector of predicted values. 
Note
See the vignette on population prediction for more details.
Author(s)
Terry Therneau
See Also
    Copyright (©) 1999–2012 R Foundation for Statistical Computing.
Licensed under the GNU General Public License.