predict-methods {aod} | R Documentation |
Methods for Function "predict" in Package "aod"
Description
“predict” methods for fitted models generated by functions in package aod.
Usage
## S4 method for signature 'glimML':
predict(object, newdata = NULL,
type = c("response", "link"), se.fit = FALSE, ...)
## S4 method for signature 'glimQL':
predict(object, newdata = NULL,
type = c("response", "link"), se.fit = FALSE, ...)
Arguments
object |
A fitted model of formal class “glimML” (functions betabin or negbin ) or
“glimQL” (functions quasibin or quasipois ). |
newdata |
A data.frame providing all the explanatory variables necessary for predictions. |
type |
A character string indicating the scale on which predictions are made: either “response” for
predictions on the observation scale, or “link” for predictions on the scale of the link (logit, cloglog
or log, depending on the call of the fitting function). |
se.fit |
A logical scalar indicating whether pointwise standard errors should be computed for the predictions. |
... |
Other arguments passed to methods. |
Methods
- object = "ANY"
- Generic function: see
predict
.
- object = "glimML"
- Compute predictions for models of formal class “glimML”, presently generated
by functions
betabin
(maximum-likelihood beta-binomial regression) and negbin
(maximum-likelihood negative-binomial regression). See the examples for these functions.
- object = "glimQL"
- Compute predictions for models of formal class “glimQL”, presently generated
by functions
quasibin
(quasi-likelihood binomial regression) and quasibin
(quasi-likelihood Poisson regression). See the examples for these functions.
See Also
predict.glm
[Package
aod version 1.1-1
Index]