lmplot: Four types of residual plots for linear models.

In HH: Statistical Analysis and Data Display: Heiberger and Holland

Four types of residual plots for linear models.

Description

Four types of residual plots for linear models. The first three are redesigns of plots that stats:::plot.lm presents. The first two show the positive residuals in col[2] and the negative residuals in color col[1]. The third and fourth use color col[1]. The fourth is based on an S-Plus panel that R\ doesn't provide.

Usage

lmplot(lm.object, ..., main=NULL,
       col=trellis.par.get("superpose.symbol")$col[1:2],
       ylim=NULL)

Arguments

lm.object	Linear model object. See lm for details.
col	Vector of color names. Only the first two are used. If not specified, then trellis.par.get("superpose.symbol")$col[1:2] is used.
main	standard main title for plots.
ylim	standard lattice argument. It is used as specified for the residVSfitted, diagQQ, and diagplot5new plots. For the scaleLocation plot, the ylim is modified to c(0, max(abs(ylim))). The main reason for using the ylim argument is to allow visual comparison of the residuals for two different models on the same scale.
...	Other arguments, currently ignored.

Details

The trellis plots from the four functions residVSfitted, scaleLocation, diagQQ, diagplot5new are displayed on the current device in a coordinated display.

Value

A list of three trellis objects is returned invisibly, the first contains the result of residVSfitted and scaleLocation together. The second diagQQ, and the third diagplot5new.

Author(s)

Richard M. Heiberger <[e-mail]>

See Also

residVSfitted, scaleLocation, diagQQ, diagplot5new.

Examples

tmp <- data.frame(y=rnorm(100), x1=rnorm(100), x2=rnorm(100))
tmp.lm <- lm(y ~ x1 + x2, data=tmp)
lmplot(tmp.lm)

HH documentation built on May 29, 2024, 6:24 a.m.