Plot diagnostics for a stability object — plot.stability • threshr

plot method for objects of class "stability" returned from stability

Usage

# S3 method for class 'stability'
plot(
  x,
  y,
  ...,
  prob = TRUE,
  top_scale = c("none", "excesses", "opposite"),
  vertical = TRUE
)

Arguments

x: an object of class "stability", a result of a call to stability.
y: Not used.
...: Additional arguments passed on to matplot, axis and/or segments.
prob: A logical scalar. If TRUE then the levels of thresholds on the lower horizontal axis are represented by the proportion of observations that lie below a threshold. If prob = FALSE then the values of the thresholds are used.
top_scale: A character scalar. If top_scale = "none" then no axis labels appear on the upper horizontal axis. If top_scale = "excesses" then the number of threshold excesses at each threshold are indicated. If top_scale = "opposite" then the type of threshold level not chosen using prob is indicated.
vertical: A logical scalar. Should the confidence intervals be depicted using a vertical line for each threshold (TRUE) or by joining up confidence limits across thresholds (FALSE)?

Value

In addition to producing the plot a list of the arguments used by matplot, axis is returned (invisibly).

Details

Produces a simple threshold diagnostic plot based on the object returned from stability. The MLEs of the GP shape parameter $$\xi$$ and approximate conf% confidence intervals for $\xi$ are plotted against the threshold used to fit the GP model. This plot is used to choose a threshold above which the underlying GP shape parameter may be approximately constant. See Chapter 4 of Coles (2001). See also the vignette "Introducing threshr". as described in . See also the vignette "Introducing threshr".

Examples

u_vec_gom <- quantile(gom, probs = seq(0, 0.9, by = 0.05))
gom_stab <- stability(data = gom, u_vec = u_vec_gom)
plot(gom_stab)