stepInformation: Stepwise point process model selection

In ppstat: Point Process Statistics

Description

Performs a stepwise, greedy model selection using an information criteria.

Usage

## S4 method for signature 'PointProcessModel'
stepInformation(model, scope = ~ 0, direction = "both", trace = 1,
steps = 1000, warmStart = TRUE, k = 2, ...)

## S4 method for signature 'MultivariatePointProcess'
stepInformation(model, scope = ~ 0, direction = "both", trace = 1,
steps = 1000, warmStart = TRUE, k = 2, ...)

## S4 method for signature 'PointProcessModel'
getInformation(model, k = 2, ...)

Arguments

model	a "PointProcessModel" object. This model is used as the initial model in the stepwise search.
scope	a formula. Only the right hand side matters. It specifies an upper bound of models to consider. The default, ~ 0, results in the upper bound being the formula of model. If specified, the upper bound is the model whose formula has a right hand side, which is the sum of terms from scope and the formula of model.

direction	a character, one of "both" or "backward", with a default of "both". Determines the direction of the search.
trace	a numeric. If set to 1 (the default) the algorithm prints tabular information while running. A value of 2 gives a different format for the information and the selected model in each step is printed. Other values result in no printed information.

steps	a numeric. The maximum number of steps. Default value is 1000.
warmStart	a logical. Determines if any updated fit should be computed with initial parameter values taken from the previous fit. The default is TRUE.
k	a numeric. The multiple of the number of degrees of freedom used for the penalty. The default is k = 2, which gives genuine AIC.
...	other arguments

Details

N.B. The function stepAIC from MASS was used as a template for this function. Several of the functionalities of stepAIC are currently not implemented in this method.

Value

The stepwise selection results in a model for which the information criteria can not be reduced further by additional steps (a local minimizer). This resulting model, an object of class "PointProcessModel", is returned (invisibly) from stepInformation.

For getInformation a numeric is returned.

See Also

stepAIC, pointProcessModel.

ppstat documentation built on May 2, 2019, 5:26 p.m.