revdbayes (version 1.5.5)

plot.evpred: Plot diagnostics for an evpred object

Description

plot method for class "evpred". Plots summarising the predictive distribution of the largest value to be observed in N years are produced. The plot produced depends on x$type. If x$type = "d", "p" or "q" then matplot is used to produce a line plot of the predictive density, distribution or quantile function, respectively, with a line for each value of N in x$n_years. If x$type = "r" then estimates of the predictive density (from density) are plotted with a line for each N. If x$type = "i" then lines representing estimated predictive intervals are plotted, with the level of the interval indicated next to the line.

Usage

# S3 method for evpred
plot(
  x,
  ...,
  leg_pos = NULL,
  leg_text = NULL,
  which_int = c("long", "short", "both")
)

Value

Nothing is returned.

Arguments

x

An object of class "evpost", a result of a call to rpost.

...

Additional arguments passed on to matplot.

leg_pos

A character scalar. Keyword for the position of legend. See legend.

leg_text

A character or expression vector. Text for legend. See legend.

which_int

A character scalar. If x$type = "i" which intervals should be plotted? "long" for equi-tailed intervals, "short" for the shortest possible intervals, "both" for both.

See Also

predict.evpost for the S3 predict method for objects of class evpost.

Examples

data(portpirie)
mat <- diag(c(10000, 10000, 100))
pn <- set_prior(prior = "norm", model = "gev", mean = c(0,0,0), cov = mat)
gevp  <- rpost(n = 1000, model = "gev", prior = pn, data = portpirie)

# Predictive density function
d_gevp <- predict(gevp, type = "d", n_years = c(100, 1000))
plot(d_gevp)

# Predictive distribution function
p_gevp <- predict(gevp, type = "p", n_years = c(100, 1000))
plot(p_gevp)

# Predictive quantiles
q_gevp <- predict(gevp, type = "q", n_years = c(100, 1000))
plot(q_gevp)

# Predictive intervals
i_gevp <- predict(gevp, type = "i", n_years = c(100, 1000), hpd = TRUE)
plot(i_gevp, which_int = "both")

# Sample from predictive distribution
r_gevp <- predict(gevp, type = "r", n_years = c(100, 1000))
plot(r_gevp)
plot(r_gevp, xlim = c(4, 10))