count.fit: Fits four different count models and compares them.

In catregs: Post-Estimation Functions for Generalized Linear Mixed Models

Fits four different count models and compares them.

Description

Given a Poisson model object, count.fit fits Poisson, negative binomial, zero-inflated Poisson, and zero-inflated negative binomial models to the data. It reports results of Vuong tests between the zero-inflated and non-zer-inflated models, summarizes the information criteria of the four models, summarizes the model output of the four models, creates a ggplot object of coefficient plots for each model, and creates a ggplot object of model residuals.

Usage

count.fit(m1,y.range,rounded=3,use.color="yes")

Arguments

m1	A Poisson regression model, as estimated via the glm function.
y.range	The observed response range for the count outcome. For example, if the observed range is 0 to 18, this would be 0:18
rounded	The number of decimal places to round the output. The default is 3.
use.color	Whether to use color in the ggplot objects. Default is "yes"

Value

ic	A data.frame summarizing the information criteria for the four models. Bayesian and Akaike's informaiton criteria are included.
models	A summary of the model estimates, including coefficients and standard errors.
pic	A ggplot object illustrating model residuals for each type of model.
models.pic	A ggplot object of coefficient plots from each type of model.

Author(s)

David Melamed

Examples

data("LF06art")
p1 <- glm(art ~ fem + mar + kid5 + phd + ment , family = "poisson", data = LF06art)
table(LF06art$art)
fit<-count.fit(p1,0:19)
names(fit)

catregs documentation built on Sept. 11, 2024, 6:18 p.m.