plot.wle.lm {wle} | R Documentation |
The plot.wle.lm
function plots a separate graph
windows for each root. In each windows four plots are printed:
residuals vs fitted, normal qq plot of the residuals, weighted
residuals vs weighted fitted, normal qq plot of the weighted
residuals. A summary plot is also printed: in the diagonal, the value
of the weigths vs position of the observations for each root; in the
upper diagonal residuals vs residuals of two different roots; in the
lower diagonal weights vs weights of two different roots. The roots
and the graphs can be chosen by the arguments roots
,
which.main
and which
.
## S3 method for class 'wle.lm': plot(x, roots, which=1:4, which.main, level.weight=0.5, ask = dev.interactive(), col=c(2, 1, 3), id.n=3, labels.id, cex.id = 0.75, verbose=FALSE, ...)
x |
an object of class wle.lm . |
roots |
a vector specify for which roots the plots are required. |
which |
if a subset of the plots for each root is required, specify a subset of the numbers 0:4 , 0 means no plots. |
which.main |
if a subset of the plots for the main graphic is
required, specify a subset of the numbers 0:roots^2 , 0 means
no plots. The plots are specified by columns. |
level.weight |
value of the weight under which an observations is marked with different color. |
ask |
logical; if TRUE , the user is asked before each plot, see par(ask=.) . |
col |
a vector of 3 elements, to specify colors for the plots. |
id.n |
number of points to be labelled in some plots, starting with the ones with less weight. |
labels.id |
vector of labels, from which the labels for less weighted points will be chosen. If missing uses observation numbers. |
cex.id |
magnification of point labels. |
verbose |
if TRUE warnings are printed. |
... |
graphical parameters can be given as arguments. |
Claudio Agostinelli
wle.lm
a function for estimating linear models with normal distribution error and normal kernel.
library(wle) data(artificial) result <- wle.lm(y~x1+x2+x3, data=artificial, boot=40, group=6, num.sol=2) result plot(result) # all plots, default behavior plot(result, roots=1) # only first root, one plot for window par(mfcol=c(2,2)) plot(result, roots=1) # only first root, as usual plot(result, roots=2, which=1, which.main=0) # only second root, only residual vs fitted values plot plot(result, which=1) # main plot + residual vs fitted values plot for each root par(mfcol=c(3,2)) plot(result, which=1) # main plot + residual vs fitted values plot for each root all in the same window