Diagnostic Plots for Robustly Fitted Generalized Linear Models
DESCRIPTION:
Creates a set plots useful for assessing a fitted generalized linear model
of class glmRob. 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.
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.
DETAILS:
The glmRob object is coerced to an glmfm object containing a single
model. The plotting is then accomplished by calling plot.glmfm.
REFERENCES:
Garcia-Ben, M. and Yohai, V. J. (2000). Quantile-quantile
plot for the deviance residuals in the Generalized
Linear Model. Technical Report. Department of Mathematics,
FCEN,
University of Buenos Aires.