denpro (version 0.9.2)

shape2d: Returns a 2D volume function or 2D probability content function

Description

Calculates a 2D volume function or 2D probability content function from a sequence of shape trees, corresponding to a sequence of level sets of a multivariate function to be visualized. A 2D volume function and 2D probability content function are visualizations of a multivariate density with a 2D function. One joins a series of radius functions or tail probability functions to get a 2D visualization. Radius functions and tail probability functions are 1D functions which visualize the shape of level sets of a multivariate density.

Usage

shape2d(shtseq, gnum = 500, type = "radius", type2 = "slice", 
gnum2 = 1000, ngrid = 30, norma = FALSE, xmax = 10, modelim = 2, 
exmalim = NULL, maxnum = NULL)

Arguments

shtseq

A list of shape trees, made by function "stseq"; the shape trees in the list correspond to a grid of level sets of the function to be visualized

gnum

Number of grid points in the radius function or in the tail probability function (used when one transforms a shape tree to a 1D function)

type

"radius" or "proba"; whether a 2D volume function or 2D probability content function will be calculated

type2

"slice" or "boundary"; whether the 1D functions are slices or level sets of the 2D function. The option "boundary" is not recommended at the moment

gnum2

when type2="boundary", then gnum*gnum is the size of the grid of the 2D function, and "gnum2" is the grid used for transforming shape trees to 1D functions

ngrid

positive integer; one prunes the 1D functions to have "ngrid" level sets

norma

TRUE if one uses dimension normalization for the volumes; then volumes are tranformed to (volume/V)pow(1/d), where V is the volume of the d-dimensional unit ball

xmax

internal

modelim

internal

exmalim

internal

maxnum

internal

Value

A list containing elements x, y, z, whose meaning is the same as in function "persp".

References

Jussi Klemela (2005). Visualization of the spread of multivariate distributions.

See Also

stseq, plotvolu2d

Examples

N<-c(30,30)  
lnum<-20
func<-"gumbel"
marginal<-"normal"
g<-2
ver<-3
support<-c(-ver,ver,-ver,ver)
st<-stseq(N,lnum,func=func,marginal=marginal,g=g,support=support)

gnum<-50 
ngrid=50 
vd<-shape2d(st,gnum=gnum,ngrid=ngrid) 

plotvolu2d(vd)

type<-"proba"
vd3<-shape2d(st,gnum=gnum,type=type,ngrid=ngrid,norma=TRUE)

plotvolu2d(vd3)