Diagnostic Plots for an "lms" Object

DESCRIPTION:

Creates a set of plots suitable for assessing a fitted linear model of class lms.

USAGE:

plot.lms(lms.obj, ask=F, id.n=3, ...)

REQUIRED ARGUMENTS:

lms.obj
an lms object.

OPTIONAL ARGUMENTS:

ask
if TRUE, plot.lms operates in interactive mode.
id.n
number of points (must be less than the number of observations) to be identified in the appropriate plots. These will be the id.n most extreme points. Set to FALSE if no points are to be identified. Default: 3.

SIDE EFFECTS:

An appropriate x-y plot is produced to display diagnostic plots. These can be one or all of the following choices:

Standardized LMS Residuals vs Fitted Values

Index Plot of Standardized LMS Residuals
Normal QQplot of Residuals
Diagnostic plot

Each of these plots can be adorned with labels for the id.n most extreme points, depending on the options chosen.

DETAILS:

This function is a method for the generic function for class lms. It can be invoked by calling for an object x of the appropriate class, or directly by calling regardless of the class of the object.

When ask=T, rather than produce each plot sequentially, plot.lms displays a menu listing all the plots that can be produced. If the menu is not desired but a pause between plots is still wanted one must set par(ask=T) before invoking this command with argument ask=F.

The last plot - the Diagnostic plot - summarizes the results obtained from lmsreg and cov.mve. The plot is only produced if the option mve=T has been set when calling lmsreg. The diagnostic plot shows the standardized LMS residuals versus the robust distances (based on MVE) of the corresponding x. This plot allows the user to distinguish between regular observations, vertical outliers, good leverage points, and bad leverage points.

SEE ALSO:

, , .

EXAMPLES:

lms.obj <- lmsreg(stack.x, stack.loss)
plot(lms.obj, ask=T) # interactive version
oldpar <- par(ask=T)
plot(lms.obj)
plot(lms.obj, id.n=10)
par(oldpar)  # reset ask par