plotting {ipptoolbox} | R Documentation |
Various routines to visualize BPAs
dscdf(x, xrange = NULL, col = c(3, 4), ..., newplot = TRUE) dsqqplot(ds, sample, points = FALSE) dsbel(x) dspl(x)
x |
BPA to plot |
xrange |
Optional: xrange=c(1,5) limits the x-axis to [1,5] |
col |
Optional: col = c(3,4) defines the colors of the Bel and the Pl line. |
... |
Optional: additional parameters passed to the plot routine (except linewidth, linetype and color) |
newplot |
Optional: Create a new plot or draw lines onto an existing plot. |
sample |
set of data values, e. g. c(1.5,2,3.5) |
ds |
BPA for qq plot resembling sample |
points |
Optional, points=TRUE: plot the sample values as points + line, otherwise only a straight line. |
The function dscdf plots Bel([-Inf,x]) and Pl([-Inf,x]), the bounds on all CDFs. Arbitrary graphics parameters can be passed to dscdf. dsqqplot plots a quantile-quantile plot between data and a BPA. dsbel and dspl return the function Bel([-Inf,x])/ Pl([-Inf,x]) as a set of (x,y)-type points, e. g. for own plotting routines.
The example generates a BPA. It plots the BPA first without, then with additional parameters. Finally it uses the functions dsbel and dspl for plotting points into the diagram. Then it generates a BPA from a normal pbox with imprecise standard deviation. It tests with a QQ plot, if the function fits well to a random sample from N(0,1).
dscdf returns the x and y coordinates of the plotted lines for further use.
Philipp Limbourg <p.limbourg@uni-due.de>
print("Plot new BPA") a=dsstruct(rbind(c(1,2,0.3),c(1.5,5,0.5),c(4,6,0.2))) dscdf(a) dscdf(a,xrange=c(0,7),xlab="Important parameter") print("Bel([-Inf,x])") bel=dsbel(a) print("Pl([-Inf,x])") pl=dspl(a) points(pl,lwd=5,col='red') points(bel,lwd=5,col='pink') print("See if a pbox from N(0,[0.5,1.5]) fits data from N(0,1)") data=rnorm(100,0,1) pbox=dsodf(qnorm,1000,0,dsstruct(c(0.5,1.5,1))) print("Plot pbox") dscdf(pbox) print("qq plot") dsqqplot(pbox,data)