Plot the solution path

Plots each off-diagonal precision entry (or, with normalize = TRUE, partial correlation) against v0 along the ladder, one panel per group. Entries whose PIP is below thres at a step are drawn as 0.

Usage

plot_path(
  v0s,
  obj,
  thres = 0.5,
  normalize = FALSE,
  xlab = "",
  ylab = "",
  main = "",
  par = c(1, 2),
  ylim = NULL,
  reverse = FALSE,
  position = "bottomright",
  ...
)

Arguments

v0s

The v0 values of the fit (obj$v0s).

obj

An ssjgl object.

thres

PIP threshold. Default 0.5.

normalize

Logical. If TRUE, plot partial correlations instead of precision entries. Default FALSE.

xlab

String for xlab. Default empty.

ylab

string for y-axis label. Default empty.

main

string for main title. Default empty.

par

Layout c(rows, cols) for the K panels, passed to par(mfrow = ) and restored on exit. Default c(1, 2); NULL leaves the layout alone.

ylim

default NULL and will calculate from range

reverse

Default FALSE. If want to reverse direction of plot.

position

Legend position, or "none". Default "bottomright".

...

other parameters passed in to base plot

Value

Invisible NULL. Called for side effect (plot).

Examples

sim <- simulate_ssjgl_data(K = 2, p = 8, n = 100, seed = 42)
fit <- ssjgl(sim$data_list, lambda0 = 1, lambda1 = 1, lambda2 = 1,
             v0s = make_v0_ladder(1, n_steps = 5, min_ratio = 5, max_ratio = 200),
             standardize = TRUE, maxitr.em = 50, impute = FALSE)
plot_path(fit$v0s, fit, thres = 0.5,
          xlab = expression(v[0]), ylab = "Precision entries",
          main = c("Group 1", "Group 2"))