image.lmSubsets: Heatmaps for Visualizing Subset Selections

In lmSubsets: Exact linear subset regression

Description

Visualizations of best subset selections carried out via lmSubsets.

Usage

## S3 method for class 'lmSubsets'
image(x, best = 1, size = NULL, which = NULL,
  hilite = "BIC", uline = hilite,
  main = "Subset selection", xlab = "", ylab = NULL,
  col = gray.colors(1), tint = 0.8, gaps = "white",
  hilite.col = "red", hilite.tint = tint,
  xaxs = "i", yaxs = "i", cex = 0.9, srt = 45, ...)

Arguments

x	An object of class lmSubsets.
best	Rank of best subset solution.
size	Subset sizes to be plotted (default: all).
which	Regressors to be plotted (default: all).
hilite, uline	Logical, character, or numeric specifying which subset size(s) should be highlighted or underlined, respectively.
main, xlab, ylab	Main and axis titles.
col, tint	Color and tint for selected and non-selected regressors, respectively.
gaps	Color of border between tiles.
hilite.col, hilite.tint	Color and tint for highlighted subset sizes.
xaxs, yaxs	The style of axis interval calculation to be used (see par for details).
cex	Expansion factor for labels.
srt	Angle of the x-axis labels.
...	Further arguments passed to plot.

See Also

lmSubsets

Examples

## all-subsets regression
data("AirPollution", package = "lmSubsets")
lms <- lmSubsets(mortality ~ ., data = AirPollution)

## visualize the selected subsets for all sizes and
## highlight best BIC solution (default)
image(lms)

## variations
## only some subset sizes and no highlighting
image(lms, size = 3:10, hilite = FALSE)
## highlight multiple sizes
image(lms, hilite = c(5, 10), hilite.col = c("green", "blue"))
## highlight subset size 3 but do not underline (and vice versa)
image(lms, hilite = 3, uline = FALSE)
image(lms, hilite = FALSE, uline = 3)

lmSubsets documentation built on May 31, 2017, 3:55 a.m.