rgrs: Relaxed Guided Random Search

In Datasmith: Tools to Complete Euclidean Distance Matrices

Description

rgrs Produce a point configuration given the edge lengths of the desired minimum spanning tree

Usage

rgrs(edges = NULL, d, n = NULL, theta = NULL, outlying = "N",
  skew = "N", stringy = "N")

Arguments

edges	A numeric vector containing the desired edge lengths of the minimum spanning tree. If n is specified, must be NULL.
d	the dimension of the resulting configuration.
n	the desired number of edge lengths to simulate. If edges is specified, must be set to NULL.
theta	Angle restriction during point proposal of the form (theta1,theta2,p), where p represents the probability of confining the proposal to [theta1,theta2]. Only used for d=2, otherwise NULL. See details for more in depth explanation.
outlying	One of "L", "M", or "H", specifying if the simulated edge lengths should have a Low, Medium, or High outlying scagnostic value.
skew	One of "L", "M", or "H", specifying if the simulated edge lengths should have a Low, Medium, or High skew scagnostic value.
stringy	One of "L", "M", or "H", specifying if the simulated edge lengths should have a Low, Medium, or High stringy scagnostic value. A numeric scalar specifying a value of stringy is also accepted.

Details

In 2-dimensions, when a new point is proposed, the position for the new point is determined by:

x <- x0 + r*sin(theta) y <- y0 + r*cos(theta)

where (x0,y0) is the base point, and r is the minimum spanning tree distance. theta is generated from a uniform distribution on (-pi,pi). By specifying the theta argument, the proposed theta is restricted, and is then generated from Uniform(theta1,theta2) or Uniform(-theta2,-theta1) with equal probability. This restriction allows the user to introduce striation into their point configuration.

Value

An nxd matrix containing the d-dimensional locations of the points.

Examples

#An example where edge lengths are supplied
EL <- runif(100,0,1)
rgrs(edges = EL, d = 2)
rgrs(edges = EL, d = 3) 
 
#An Example where edge lengths are simulated internally
rgrs(d=2, n=100)
rgrs(d=3, n=100)
rgrs(d=2, n=100, outlying="H")
rgrs(d=2, n=100, skew = "M")
rgrs(d=2, n=100, stringy = "H")
 
#An Example making use of theta
rgrs(d=2, n=100, theta=c(pi/4,pi/3,.5))
 

Datasmith documentation built on May 29, 2017, 10:19 p.m.