R: Predict method for GeDS objects

predict.GeDS {GeDS}

R Documentation

Predict method for GeDS objects

Description

This is a user friendly method to compute predictions from GeDS objects.

Usage

## S3 method for class 'GeDS'
predict(object, newdata, type = c("response", "link", "terms"), n = 3L, ...)

Arguments

`object`	the `GeDS-class` object for which the computation of the predicted values is required.
`newdata`	an optional `data.frame`, `list` or `environment` containing values of the independent variables for which predicted values of the predictor model (including the GeDS and the parametric components) should be computed. If left empty the values are extracted from the object `x` itself.
`type`	character string indicating the type of prediction required. By default it is equal to `"response"`, i.e. the result is on the scale of the response variable. See details for the other options.
`n`	integer value (2, 3 or 4) specifying the order (`=` degree `+ 1`) of the GeDS fit whose predicted values should be computed. By default equal to `3L`. Non-integer values will be passed to the function `as.integer`.
`...`	potentially further arguments (required by the definition of the generic function). They are ignored, but with a warning.

Details

This is a method for the function predict that allows the user to handle GeDS-Class objects.

In analogy with the function predict.glm in the stats package, the user can specify the scale on which the predictions should be computed through the argument type. If the predictions are required to be on the scale of the response variable, the user should set type = "response", which is the default. Alternatively if one wants the predictions to be on the predictor scale, it is necessary to set type = "link". By specifying type = "terms", it is possible to inspect the predicted values separately for each single independent variable which enter either the GeD spline component or the parametric component of the predictor model. In this case the returned result is a matrix whose columns correspond to the terms supplied via newdata or extracted from the object.

As GeDS objects contain three different fits (linear, quadratic and cubic), it is possible to specify the order for which GeDS predictions are required via the input argument n.

Value