summary.glm
Summary Method for Fitted Generalized Linear Models

Description

Returns a summary list for fitted generalized linear models.

Usage

summary.glm(object, dispersion = NULL, correlation = FALSE, 
    symbolic.cor = FALSE, ...)
print(x, digits = max(3, getOption("digits") - 3), 
    symbolic.cor = x$symbolic.cor, 
    signif.stars = getOption("show.signif.stars"), ...)

Arguments

object a fitted glm model object. Since this function expects the components returned by the glm() function, the object that you specify should be the result of some fit that inherited from the "glm" class.
dispersion a numeric value that specifies the dispersion parameter. For the poisson or binomial families, the default is 1. For other families, the default is the residual Chi-square statistic divided by the residual degrees of freedom. For example, in the Gaussian family the dispersion parameter is the error variance and the default is the Residual sum of squares divided by the residual degrees of freedom. If you specify a value of 0, the Chi-squared estimate is used in all cases.
correlation a logical value. If FALSE, the function does not compute the correlation matrix for the coefficients. This is useful for models with many coefficients.
symbolic.cor a logical value. In summary.glm, if correlation = TRUE and object$rank > 0, this argument retains the assigned value and is included in the summary.
x an object of class summary.glm.
digits an integer used in print, to indicate the number of significant digits to use when printing.
signif.stars a logical value. If TRUE, significance stars are displayed along with the coefficient.
dots extra arguments passed to or from other methods.

Details

This function is a method for the generic function summary for class glm. It can be invoked by callingsummary for an object x of the appropriate class, or directly by calling summary.glm regardless of the class of the object.
Although some people call this function to print linear coefficients and their standard errors for gam models, because summary.glm ignores any colinearity between the nonlinear part of gam smooths and other terms, this can result in a drastic under-estimate of the standard errors.
Value
returns a list with the following components.
callelement is returned as it is contained in the object.
termselement is returned as it is contained in the object.
familyelement is returned as it is contained in the object.
devianceelement is returned as it is contained in the object.
aicelement is returned as it is contained in the object.
contrastselement is returned as it is contained in the object.
df.residualelement is returned as it is contained in the object.
null.devianceelement is returned as it is contained in the object.
df.nullelement is returned as it is contained in the object.
iterelement is returned as it is contained in the object.
na.actionelement is returned as it is contained in the object.
deviance.resid the deviance residuals, as produced by residuals(object).
coefficients a matrix with four columns that contain the coefficients, their standard errors, z-values, and p-values.
aliased a logical vector that indicates if the original coefficients are aliased.
dispersion a numeric value that reports the dispersion parameter used in the dispersion argument.
df a numeric vector with length = 3 that contains the rank of the model, number of residual degrees of freedom, and number of non-aliased coefficients.
cov.unscaled the unscaled covariance matrix. That is, a matrix such that if you multiplied the matrix by the dispersion parameter, or estimate thereof, an estimated (asymptotic) covariance matrix for the coefficients would be produced.
cov.scaled the covariance matrix scaled by dispersion. If object$rank, the value is dispersion * covmat.unscaled, otherwise cov.scaled is the same as cov.unscaled.
correlation the computed correlation matrix for the coefficients in the model. Exists only when argument correlation = TRUE and object$rank > 0.
symbolic.cor a logical value that reports the value assigned to the symbolic.cor argument. Exists only when argument correlation = TRUE and object$rank > 0.
See Also
summary, summary.gam, glm.object, print
Package stats version 6.1.5-18
Package Index