predict.hurdle: Methods for hurdle Objects

Description

Methods for extracting information from fitted hurdle regression model objects of class "hurdle".

Usage

# S3 method for hurdle
predict(object, newdata,
  type = c("response", "prob", "count", "zero"), na.action = na.pass,
  at = NULL, …)
# S3 method for hurdle
residuals(object, type = c("pearson", "response"), …)
# S3 method for hurdle
coef(object, model = c("full", "count", "zero"), …)
# S3 method for hurdle
vcov(object, model = c("full", "count", "zero"), …)
# S3 method for hurdle
terms(x, model = c("count", "zero"), …)
# S3 method for hurdle
model.matrix(object, model = c("count", "zero"), …)

Arguments

object, x

an object of class "hurdle" as returned by hurdle.

newdata

optionally, a data frame in which to look for variables with which to predict. If omitted, the original observations are used.

type

character specifying the type of predictions or residuals, respectively. For details see below.

na.action

function determining what should be done with missing values in newdata. The default is to predict NA.

optionally, if type = "prob", a numeric vector at which the probabilities are evaluated. By default 0:max(y) is used where y is the original observed response.

model

character specifying for which component of the model the terms or model matrix should be extracted.

…

currently not used.

Details

A set of standard extractor functions for fitted model objects is available for objects of class "hurdle", including methods to the generic functions print and summary which print the estimated coefficients along with some further information. The summary in particular supplies partial Wald tests based on the coefficients and the covariance matrix (estimated from the Hessian in the numerical optimization of the log-likelihood). As usual, the summary method returns an object of class "summary.hurdle" containing the relevant summary statistics which can subsequently be printed using the associated print method.

The methods for coef and vcov by default return a single vector of coefficients and their associated covariance matrix, respectively, i.e., all coefficients are concatenated. By setting the model argument, the estimates for the corresponding model component can be extracted.

Both the fitted and predict methods can compute fitted responses. The latter additionally provides the predicted density (i.e., probabilities for the observed counts), the predicted mean from the count component (without zero hurdle) and the predicted ratio of probabilities for observing a non-zero count. The latter is the ratio of probabilities for a non-zero implied by the zero hurdle component and a non-zero count in the non-truncated count distribution. See also Appendix C in Zeileis et al. (2008).

The residuals method can compute raw residuals (observed - fitted) and Pearson residuals (raw residuals scaled by square root of variance function).

The terms and model.matrix extractors can be used to extract the relevant information for either component of the model.

A logLik method is provided, hence AIC can be called to compute information criteria.

References

Zeileis, Achim, Christian Kleiber and Simon Jackman 2008. “Regression Models for Count Data in R.” Journal of Statistical Software, 27(8). URL http://www.jstatsoft.org/v27/i08/.

Examples

Run this code

# NOT RUN {
data("bioChemists", package = "pscl")
fm <- hurdle(art ~ ., data = bioChemists)

plot(residuals(fm) ~ fitted(fm))

coef(fm)
coef(fm, model = "zero")

summary(fm)
logLik(fm)
# }