predictCounts {WALS}

Internal methods: Predict probability for counts

Description

Predicts the probability of counts given a family object of class "familyWALScount". Only works for count data models.

Usage

predictCounts(x, ...)

## S3 method for class 'familyWALScount'
predictCounts(x, yUnique, rowNames, eta, ...)

Arguments

x

object of class "familyWALScount".

...

Further parameters passed to density() function in family.

yUnique

vector. The counts (larger or equal to zero) which to predict probabilities for.

rowNames

vector. The names of the observations.

eta

vector. The fitted linear link \hat{\eta} of the model.

Details

"familyWALScount" objects are used in the fitting methods walsNB, walsNBmatrix, walsGLM or walsGLMmatrix. For the latter two, only the family poissonWALS is currently supported.

predictCounts() is not available for objects of any class except for "familyWALScount".

The predictCounts.familyWALScount() method is a modified version of the predict.hurdle() method from the countreg package version 0.2-1 (2023-06-13) (Zeileis and Kleiber 2023; Zeileis et al. 2008) using the argument type = "prob".

Value

Returns a matrix of dimension length(eta) times length{yUnique} with the predicted probabilities of the counts given in yUnique for every observation in eta.

References

Zeileis A, Kleiber C (2023). countreg: Count Data Regression. R package version 0.2-1, https://r-forge.r-project.org/projects/countreg/.

Zeileis A, Kleiber C, Jackman S (2008). “Regression Models for Count Data in R.” Journal of Statistical Software, 27(8), 1–25. doi:10.18637/jss.v027.i08.