Diagnostic Plots for an "lts" Object

DESCRIPTION:

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

USAGE:

plot.lts(x, ask=F, id.n=3, which.plots=NULL, ...) 

REQUIRED ARGUMENTS:

x
an object of class "lts".

OPTIONAL ARGUMENTS:

ask
if TRUE, plot.lts 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.
which.plots
Integer vector indicating which plots to display.

SIDE EFFECTS:

Bivariate plots are produced to display diagnostic information. These can be one or all of the following choices:

Standardized LTS Residuals versus LTS Fitted Values

Index Plot of Standardized LTS Residuals
Normal QQ Plot of LTS Residuals
Diagnostic plot of LTS residuals vs Robust Distances of x-rows

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 lts. 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.lts 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 the plot command.

The last plot is the diagnostic plot of Rousseeuw and van Zomeren (1990). It summarizes the results obtained from ltsreg and cov.mcd. The plot is only produced if the option mcd = T has been set when calling ltsreg. The diagnostic plot shows the standardized LTS residuals versus the robust distances (based on MCD) of the corresponding x. This plot allows the user to distinguish between regular observations, vertical outliers, good leverage points, and bad leverage points.

REFERENCES:

Rousseeuw, P. J. and van Zomeren, B. C. (1990). Unmasking multivariate outliers and leverage points. Journal of the American Statistical Association, 85, 633-639.

SEE ALSO:

, , , .

EXAMPLES:

lts.obj <- ltsreg(stack.x, stack.loss) 
plot(lts.obj, ask=T)  # interactive version 
oldpar <- par(ask=T) 
plot(lts.obj) 
plot(lts.obj, id.n=10) 
par(oldpar)  # restore ask par