Interface GlmResult

AIC; NaN for the quasi families.

Whether the step was truncated (halved) in the last iteration.

Coefficients (intercept first); NaN for aliased columns.

Whether the iterations converged.

Residual deviance.

Degrees of freedom of the null deviance (R's df.null).

Residual degrees of freedom (R's df.residual).

The family's name and the link's name ("mu^0.333" for a power link).

Fitted means, one per row (rows with a missing value dropped).

Number of iterations.

Linear predictors, one per row, offset included (R's linear.predictors).

The link's name, such as "logit".

The data and settings of the fit, for the refits of glmConfint and glmAnova.

Deviance of the model with the intercept only, or of the empty model without intercept (R's null.deviance).

Prior weights after the response was read (multiplied by the sizes for a two-column binomial response).

QR decomposition of the last weighted least squares problem (undefined for a model without columns).

Numerical rank of the model matrix.

Working residuals.

R's warnings, each once, in order of first appearance.

Working weights of the last least squares step (0 for rows not used).

The response as the family reads it (proportions for a two-column binomial response).