tropical.centroid.MH {TML} | R Documentation |
Centroid-based sampling using Metropolis filter
Description
This function samples points on a tropical line segment about a location parameter for a given scale parameter defined in terms of tropical distance
Usage
trop.centroid.MH(D, x0, m, s, n, I = 50, tadd = max)
trop.centroid.MH.square(D, x0, m, s, n, I = 50, tadd = max)
Arguments
D |
matrix of vertices of a tropical polytope; each row is a vertex |
x0 |
initial point for sampler, numeric vector |
m |
location parameter; numeric vector |
s |
scale parameter; scalar |
n |
number of points to sample |
I |
states in Markov chain |
tadd |
function; max indicates max-plus addition, min indicates min-plus addition. Defaults to max |
Value
matrix of n sampled points where each point is a row
Author(s)
David Barnhill david.barnhill@nps.edu
References
Yoshida, Ruriko, Keiji Miura and David Barnhill (2022). Hit and Run Sampling from Tropically Convex Sets.
Examples
D1 <-matrix(c(0,0,0,0,10,0,0,0,10),3,3,TRUE)
D2 <-matrix(c(0,10,10,0,10,0,0,0,10),3,3,TRUE)
x0 <- c(0,0,0)
m1<-c(0,5,5)
m2<-c(0,-1,1)
s<-1
n<-10
trop.centroid.MH(D1, x0, m1, s, n, I=50)
trop.centroid.MH.square(D1, x0,m1, s, n, I=50)
trop.centroid.MH(D2, x0, m1, s, n, I=50,tadd=min)
trop.centroid.MH.square(D2, x0,m2, s, n, I=50,tadd=min)
[Package TML version 2.3.0 Index]