Comparison Plots for Generalized Linear Models

DESCRIPTION:

Creates a set plots useful for comparing fitted generalized linear models stored in a class fit.models object belonging to virtual class glmfm. The plot options are (2) Deviances vs Fitted Values, (3) Response vs Fitted Values, (4) QQ-Plot of Pearson Residuals, (5) Deviances QQ-Plot, (6) Standardized Deviances vs Robust Distances, (7) Standardized Deviances vs Index (Time), and (8) Sqrt of abs(Deviances) vs Fitted Values.

USAGE:

plot.glmfm(x, which.plots="all", type="pearson", chisq.percent=0.99,
           vertical.outlier=0.99, smooths=F, rugplot=F, id.n=3, ...)

REQUIRED ARGUMENTS:

x
a fit.models object belonging to virtual class glmfm.

OPTIONAL ARGUMENTS:

which.plots
either "ask", "all", or an integer vector specifying which plots to draw. If which.plots is an integer vector, use the plot numbers given in the description above (or in the "ask" menu).
type
type of residuals, with choices "deviance", "pearson", "working" or "response"; "pearson" is the default.
chisq.percent
p-value used to calculate the chi-squared quantile used as the outlier threshold for robust distances.
vertical.outlier
p-value used to calculate the standard normal quantile used as the outlier threshold for residuals.
smooths
if TRUE, smooth curves are approximated to the scatterplots using loess.smooth and added to the appropriate plots.
rugplot
if TRUE, a univariate histogram or rugplot is displayed along the base of each plot, showing the occurrence of each x-value; ties are broken by jittering.
id.n
number of outliers identified in plots.

VALUE:

x is invisibly returned.

SIDE EFFECTS:

The selected plots are drawn on a graphics device.

SEE ALSO:

, .

EXAMPLES:

breslow.mle <- glm(sumY ~ Age10 + Base4 * Trt, family=poisson,
     data=breslow.dat)
breslow.rob <- glmRob(sumY ~ Age10 + Base4 * Trt, family=poisson,
     data=breslow.dat)
breslow.fm <- fit.models(breslow.rob, breslow.mle)
##
## breslow.fm is a fit.models object with virtual class glmfm.
##
plot(breslow.fm)