poissonreg offers a function to fit model to count data using Poisson generalized linear models or via different methods for zero-inflated Poisson (ZIP) models.
Let’s fit a model to the data from Agresti (2007) Table 7.6:
library(poissonreg)
library(tidymodels)
tidymodels_prefer()log_lin_fit <-
# Define the model
poisson_reg() %>%
# Choose an engine for fitting. The default is 'glm' so
# this next line is not strictly needed:
set_engine("glm") %>%
# Fit the model to the data:
fit(count ~ (.)^2, data = seniors)
log_lin_fit
## parsnip model object
##
##
## Call: stats::glm(formula = count ~ (.)^2, family = stats::poisson,
## data = data)
##
## Coefficients:
## (Intercept) marijuanayes
## 5.6334 -5.3090
## cigaretteyes alcoholyes
## -1.8867 0.4877
## marijuanayes:cigaretteyes marijuanayes:alcoholyes
## 2.8479 2.9860
## cigaretteyes:alcoholyes
## 2.0545
##
## Degrees of Freedom: 7 Total (i.e. Null); 1 Residual
## Null Deviance: 2851
## Residual Deviance: 0.374 AIC: 63.42
The different engines for the model that are provided by this package are:
show_engines("poisson_reg")
## # A tibble: 5 × 2
## engine mode
## <chr> <chr>
## 1 glm regression
## 2 hurdle regression
## 3 zeroinfl regression
## 4 glmnet regression
## 5 stan regression
Maintainer: Hannah Frick hannah@rstudio.com (ORCID)
Authors:
Max Kuhn max@rstudio.com (ORCID)
Other contributors:
RStudio [copyright holder, funder]
The model function works with the tidymodels infrastructure so that the model can be resampled, tuned, tided, etc.
Useful links: