verification (version 1.09)

discrimination.plot: Discrimination plot

Description

This function creates a plot of overlayed histograms. In the context of verification, this is often used to compare the distribution of event and no-event forecasts.

Usage

discrimination.plot(obs, pred, breaks = 11, main =
"Comparison of Distributions", xlim = c(0,1), leg.txt = c("Model A",
"Model B"), median = TRUE, ...)

Arguments

obs

A binary observation (coded {0, 1 } ).

pred

A vector of predictions.

breaks

Number of breaks in the x-axis of the histogram.

main

Title for plot.

xlim

Range of histogram - x axis.

leg.txt

Legend text.

median

If TRUE, the median values for the event and the no-event will be vertically plotted. This is intended as a non-paramtric indicator of the forecast's ability to discriminate. If FALSE, these vertical lines will be surpressed.

...

Additional plotting options.

Examples

#  A sample forecast.  

 a<- rnorm(100, mean = -1); b<- rnorm(100, mean = 1)

 A<- cbind(rep(0,100), pnorm(a)); B<- cbind(rep(1,100), pnorm(b)) 

 dat<- rbind(A,B); dat<- as.data.frame(dat)
 names(dat)<- c("obs", "pred")

 discrimination.plot(dat$obs, dat$pred, main = "Sample Plot")