plot.thres3 {ThresholdROC}

Thresholds and density plot (three-state setting)

Description

This function provides a graph including the three sample densities, the thresholds and their confidence intervals.

Usage

## S3 method for class 'thres3'
plot(x, bw = c("nrd0", "nrd0", "nrd0"), ci = TRUE,
  which.boot = c("norm", "perc"), col = c(1, 2, 3, 4),
  lty = c(1, 1, 1, 1, 2), lwd = c(1, 1, 1, 1),
  legend = TRUE, leg.pos = "topleft", leg.cex = 1,
  xlim = NULL, ylim = NULL,
  main = paste0("Threshold estimates", ifelse(ci, " and CIs", "")),
  xlab = "", ...)

Arguments

x

an object of class thres3.

bw

vector containing the bandwith for the first sample in the first position, the bandwith for the second sample in the second position and the bandwith for the third sample in the third position (to be passed to density()). Default, c("nrd0", "nrd0", "nrd0").

ci

should the confidence intervals be plotted? Default, TRUE. No confidence intervals will be plotted if x does not contain one (that is, x$CI is NULL).

which.boot

in case x contains confidence intervals calculated by bootstrapping, which one should be printed? The user can choose between "norm" (based on normal distribution) or "perc" (based on percentiles). Default, "norm". This argument is ignored if parametric confidence intervals were calculated.

col

a 4-dimensional vector containing:

col[1]: color for the density of the first sample

col[2]: color for the density of the second sample

col[3]: color for the density of the third sample

col[4]: color for the thresholds and their corresponding confidence intervals

Default, c(1, 2, 3, 4). If length(col) is not 4, col will be recycled.

lty

a 5-dimensional vector containing:

lty[1]: line type for the density of the first sample

lty[2]: line type for the density of the second sample

lty[3]: line type for the density of the third sample

lty[4]: line type for the thresholds

lty[5]: line type for the confidence intervals

Default, c(1, 1, 1, 1, 2). If length(lty) is not 5, lty will be recycled.

lwd

a 4-dimensional vector containing:

lwd[1]: line width for the density of the first sample

lwd[2]: line width for the density of the second sample

lwd[3]: line width for the density of the third sample

lwd[4]: line width for the thresholds and their corresponding confidence intervals

Default, c(1, 1, 1, 1). If length(lwd) is not 4, lwd will be recycled.

legend

logical asking if an automatic legend should be added to the graph. Default, TRUE.

leg.pos

position of the legend. Default, "topleft". Ignored if legend=FALSE.

leg.cex

number that reescales the size of the legend. Ignored if legend=FALSE. Default, 1.

xlim

2-dimensional vector indicating the lower and upper limits for x-axis. Default value (NULL) sets those limits automatically.

ylim

2-dimensional vector indicating the lower and upper limits for y-axis. Default value (NULL) sets those limits automatically.

main, xlab, ...

further arguments to be passed to plot().

Value

Estimates of the density functions for the three samples and vertical lines representing the thresholds and their confidence limits are drawn.

References

Skaltsa K, Jover L, Fuster D, Carrasco JL. (2012). Optimum threshold estimation based on cost function in a multistate diagnostic setting. Statistics in Medicine, 31:1098-1109.

See Also

thres3, lines.thres3

Examples

set.seed(1234)
n <- 100
k1 <- rlnorm(n)
k2 <- rnorm(n, 3, 1)
k3 <- rnorm(n, 5, 1)
rho <- c(1/3, 1/3, 1/3)

# assuming trinormality
start <- c(mean(k1), mean(k3)) 
thres <- thres3(k1, k2, k3, rho, dist1="norm", dist2="norm",
                dist3="norm", start=start, ci.method="param") 
plot(thres, leg.pos="topright")

# not assuming trinormality
thres <- thres3(k1, k2, k3, rho, dist1="lnorm", dist2="norm",
                dist3="norm", ci.method="boot") 
plot(thres, leg.pos="topright", which.boot="perc")