README.md

In rdmatheus/bergreg: BerG Regression for Count Data

bergreg: BerG Regression Model for Count Data

Rodrigo M. R. Medeiros [e-mail], IME-USP

The bergreg package (in development) provide a set of functions for a complete regression analysis of count data, in which it is assumed that the dependent variable follows a BerG distribution. The BerG regression was proposed by Bourguignon and De Medeiros (2021), and can be used to fit count data with overdispersion, equidispersion (but are not Poisson-distributed), and underdispersion.

Installation

You can install the current development version of bergreg from GitHub with:

devtools::install_github("rdmatheus/bergreg")

To run the above command, it is necessary that the devtools package is previously installed on R. If not, install it using the following command:

install.packages("devtools")

After installing the devtools package, if you are using Windows, install the most current RTools program. Finally, run the command devtools::install_github("rdmatheus/bergreg"), and then the package will be installed on your computer.

Example

This package provides a complete estimation and inference for the parameters. It also provides diagnostic plots, useful for assessing the goodness-of-fit of the model. The implementation is straightforward and similar to other popular packages, like betareg and glm, where the main function is bergreg(). Below is an example of some functions usage and available methods.

library(bergreg)

There are currently two data sets available in the bergreg package, grazing and bids. To access the documentation associated with this data, use help(grazing) and help(bids), respectively. In this example we use the grazing dataset.

# Data visualization (For description: ?grazing)

# Fit a double model (mean and dispersion) 

fit <- bergreg(birds ~ when + grazed | when + grazed, data = grazing)

## For bergreg() documentation:
?bergreg

# Print

fit
#> 
#> Call:
#> bergreg(formula = birds ~ when + grazed | when + grazed, data = grazing)
#> 
#> mean Coefficients:
#> [1]  2.1981771  0.2861585 -0.7489044
#> 
#> dispersion Coefficients:
#> [1]  2.43710069 -0.03712163 -0.44592813
#>
#> In addition, Log-lik value: -178.4034 
#> AIC: 368.8069 and BIC: 381.5697

# Summary

summary(fit)
#>
#> Call:
#> bergreg(formula = birds ~ when + grazed | when + grazed, data = grazing)
#>
#> Summary for quantile residuals:
#>        Mean Std. dev.  Skewness Kurtosis
#>    0.022816  0.977821 -0.033684 3.669114
#> 
#> ---------------------------------------------------------------
#> Mean model with log link:
#> ---------------------------------------------------------------
#> 
#> Coefficients:
#>             Estimate Std. Error z value  Pr(>|z|)    
#> (Intercept)  2.19818    0.31373  7.0066 2.441e-12 ***
#> whenAfter    0.28616    0.32352  0.8845   0.37642
#> grazedFeral -0.74890    0.32411 -2.3107   0.02085 *
#>
#> ---------------------------------------------------------------
#> Dispersion model with log link:
#> ---------------------------------------------------------------
#> 
#> Coefficients:
#>              Estimate Std. Error z value  Pr(>|z|)    
#> (Intercept)  2.437101   0.317344  7.6797 1.595e-14 ***
#> whenAfter   -0.037122   0.330650 -0.1123    0.9106
#> grazedFeral -0.445928   0.339223 -1.3146    0.1887    
#> ---
#> Signif. codes:  
#>   0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’  0.1 ‘ ’ 1
#> 
#> ----------------------------------------------------------------
#> 
#> In addition, Log-lik value: -178.4034
#> AIC: 368.8069 and BIC: 381.5697

# Test for constant dispersion

disp_test(fit)
#>            Score      Wald Lik. Ratio  Gradient
#> value  2.5735882 1.7394284  1.9181896 2.0715307
#> pvalue 0.2761547 0.4190713  0.3832396 0.3549546

# Plot

## Randomized quantile residuals
layout(matrix(c(1, 3, 5, 2, 4, 5), 3, 2))
plot(fit, which = 1:5, ask = FALSE)     
layout(1)

## Use type = "pearson" to obtain a summary for the Pearson residuals.

# Normal probability plot with simulated envelope 

envelope <- berg_envel(fit)
plot(envelope)

## Use type = "pearson" to obtain the normal probability plot with simulated
## envelope for the Pearson residuals

Currently, the methods implemented for "bergreg" objects are

methods(class = "bergreg")
#>
#> [1] AIC          coef         logLik       model.matrix plot         predict      print       
#> [8] residuals    summary      vcov        
#> see '?methods' for accessing help and source code

rdmatheus/bergreg documentation built on Dec. 22, 2021, 1:04 p.m.