modelfit function

Fit Statistics for generalized linear models

modelfit is used following a glm() or glm.nb() model to produce a list of model fit statistics.

modelfit(x)

Arguments

• x: the only argument is the name of the fitted glm or glm.nb function model

Details

modelfit is to be used as a post-estimation function, following the use of glm() or glm.nb().

Returns

• obs: number of model observatiions

• aic: AIC statistic

• xvars: number of model predictors

• rdof: residial degrees of freedom

• aic_n: AIC, 'aic'/'obs'

• ll: log-likelihood

• bic_r: BIC - Raftery parameterization

• bic_l: BIC - log-likelihood Standard definition (Stata)

• bic_qh: Hannan-Quinn IC statistic (Limdep)

References

Hilbe, J.M. (2011), Negative Binomial Regression, second edition, Cambridge University Press.

Hilbe, J.M. (2009), Logistic Regression Models, Chapman Hall/CRC

Author(s)

Joseph M. Hilbe, Arizona State University, and Jet Propulsion Laboratory, California Institute of technology

Note

modelfit.r must be loaded into memory in order to be effectve. Users may past modelfit.r into script editor to run, as well as load it.

See Also

glm, glm.nb

Examples

## Hilbe (2011), Table 9.17
library(MASS)
data(lbwgrp)
nb9_3 <- glm.nb(lowbw ~ smoke + race2 + race3 + offset(log(cases)), data=lbwgrp)
summary(nb9_3)
exp(coef(nb9_3))
modelfit(nb9_3)