predict.BranchGLM {BranchGLM}R Documentation

Predict Method for BranchGLM Objects

Description

Obtains predictions from BranchGLM objects.

Usage

## S3 method for class 'BranchGLM'
predict(
  object,
  newdata = NULL,
  offset = NULL,
  type = "response",
  na.action = na.pass,
  ...
)

Arguments

object

a BranchGLM object.

newdata

a data.frame, if not specified then the data the model was fit on is used.

offset

a numeric vector containing the offset variable, this is ignored if newdata is not supplied.

type

one of "linpreds" which is on the scale of the linear predictors or "response" which is on the scale of the response variable. If not specified, then "response" is used.

na.action

a function which indicates what should happen when the data contains NAs. The default is na.pass. This is ignored if newdata is not supplied and data isn't included in the supplied BranchGLM object.

...

further arguments passed to or from other methods.

Value

A numeric vector of predictions.

Examples

Data <- airquality

# Example without offset
Fit <- BranchGLM(Temp ~ ., data = Data, family = "gaussian", link = "identity")

## Using default na.action
predict(Fit)

## Using na.omit
predict(Fit, na.action = na.omit)

## Using new data
predict(Fit, newdata = Data[1:20, ], na.action = na.pass)

# Using offset
FitOffset <- BranchGLM(Temp ~ . - Day, data = Data, family = "gaussian", 
link = "identity", offset = Data$Day * -0.1)

## Getting predictions for data used to fit model
### Don't need to supply offset vector
predict(FitOffset)

## Getting predictions for new dataset
### Need to include new offset vector since we are 
### getting predictions for new dataset
predict(FitOffset, newdata = Data[1:20, ], offset = Data$Day[1:20] * -0.1)


[Package BranchGLM version 2.1.6 Index]