Skip to contents

Plot an LHWPR Coefficient or Focal-Selection Surface

Source: R/lhwpr_surface_plot.R
plot_lhwpr_surface.Rd

Display one fitted LHWPR slope over the Poincare-disk view of the Lorentz representation. A coefficient surface colors nonzero focal coefficients and leaves numerical zeros hollow. A selection surface shows the focal active set for a positively penalized predictor. These displays summarize penalized estimation and are not significance maps.

Usage

plot_lhwpr_surface(
  x,
  predictor,
  type = c("coefficient", "selection"),
  threshold = 1e-08,
  labels = NULL,
  node_names = NULL,
  sizes = NULL,
  title = NULL,
  point_palette = "Blue-Red 3",
  selection_colors = c(`Not selected` = "#D9D9D9", Selected = "#2166AC"),
  show_legend = TRUE,
  radial_rescale = TRUE,
  radial_quantile = 1,
  radial_power = 1,
  output_file = NULL,
  width = 7,
  height = 7,
  res = 220,
  ...
)

Arguments

x

A fitted "lhwpr" object.

predictor

Predictor name or one-based slope index. "Intercept" is available only for type = "coefficient".

type

Draw coefficient values or focal selection.

threshold

Nonnegative numerical threshold used to distinguish stored zeros from active coefficients.

labels, node_names, sizes

Optional point labels, longer names, and nonnegative point-size values; see plot_lhwr_poincare_map().

title

Optional plot title. NULL creates a descriptive default.

point_palette

HCL palette for nonzero coefficient values.

selection_colors

Two colors for not-selected and selected locations.

show_legend

Logical; show the selection legend or the filled-versus-hollow coefficient key.

radial_rescale, radial_quantile, radial_power

Display-only Poincare radial controls; they do not alter fitting distances.

output_file

Optional PNG, PDF, EPS, or PS output path.

width, height, res

Output-device dimensions in inches and resolution for raster output.

...

Additional arguments passed to plot_lhwr_poincare_map().

Value

Invisibly, a data frame containing plotted coordinates, the focal coefficient, the active indicator, predictor, threshold, and plot type.

Examples

set.seed(7)
H <- simulate_H2(35, r_max = 2)
dat <- data.frame(y = rnorm(35), x1 = rnorm(35), x2 = rnorm(35))
fit <- lhwpr(
  y ~ x1 + x2, dat, H$Z, bw = 18, lambda = c(0.2, 0.05),
  keep_path = "selected"
)
plot_lhwpr_surface(fit, "x1", type = "coefficient")
plot_lhwpr_surface(fit, "x1", type = "selection")