chull.center {qrjoint}

Fast Interior Point Center of Multivariate Data

Description

Calculates an interior point by averaging a small number of near-extreme points of the cloud.

Usage

 
chull.center(x, maxEPts = ncol(x) + 1, plot = FALSE)

Arguments

x

a matrix giving the data cloud.

maxEPts

integer giving the maximum number of (near)-extreme points to be used in averaging. Default is ncol(x)+1.

plot

logical indicating whether a pairwise scatter plot should be made

Details

Near extreme points are found in a space-filling manner by adding points with minimum residual conditional variance given points already included under a smooth GP specification. See Yang and Tokdar (2015), Section B.1. for more details.

Value

Returns an interior point of the data cloud. The positions of the near extreme points are returned as the attribute "EPts".

Examples

 
p <- 9
n <- 200
u <- runif(n)
require(splines)
x <- bs(u, df = p)
chull.center(x, plot = TRUE)