MRIaggr (version 1.1.5)

calcMultiPotential: Computation of the spatial potential

Description

Compute the regional potential of a spatial field. Call the calcMultiPotential_cpp C++ function.

Usage

calcMultiPotential(W_SR, W_LR, distance.ref, sample, coords, threshold = 0.01, nbGroup_min = 100, check.args = TRUE, verbose = TRUE)

Arguments

W_SR

The local neighborhood matrix. dgCMatrix. Should be normalized by row (i.e. rowSums(Wweight_SR)=1). REQUIRED.

W_LR

The regional neighborhood matrix. dgCMatrix. Should contain the distances between the observations (0 indicating infinite distance). REQUIRED.

distance.ref

The intervals of distance defining the several neighborhood orders in W_LR. numeric vector. REQUIRED.

sample

The probability membership to the group. numeric vector or character vector. REQUIRED.

coords

The voxel coordinates. matrix. REQUIRED.

threshold

The minimum value to consider non-negligible group membership. numeric. Default is 0.01.

nbGroup_min

The minimum group size of the spatial groups required for computing the potential. integer. Default is 100.

check.args

Should the validity of the arguments be checked ? logical.

verbose

Should the radius of the spatial groups be displayed ? logical.

Value

A numeric vector containing the regional potential.

Details

If check.args is set to TRUE, argument ccords must be a matrix (and not a data.frame) and the x slot of argument W_LR must contain categories of distances instead of distances. Distance categories must begin at 0 and end at length(distance.ref)-1.

Examples

# spatial field
## Not run: 
# n <- 30
# ## End(Not run)

G <- 3
coords <- data.frame(which(matrix(0, nrow = n * G, ncol = n * G) == 0, arr.ind = TRUE), 1)
optionsMRIaggr(quantiles.legend = FALSE, axes = FALSE, num.main = FALSE)

# neighborhood matrix
W_SR <- calcW(coords, range = sqrt(2), row.norm = TRUE)$W
W_LR <- calcW(coords, range = 10, row.norm = FALSE)$W
distance.ref <- seq(1, 10, 1)

# data
set.seed(10)
sample <- simulPotts(W = W_SR, G = G, rho = 3.5, iter_max = 500)

# 
multiplot(coords,sample$simulation[,2])

V1 <- calcMultiPotential(W_SR = W_SR, W_LR = W_LR, distance.ref = seq(1, 10, 1),
     sample = sample$simulation[,2], coords = coords, verbose = TRUE)

multiplot(coords, V1)		


#
sampleV <- (apply(sample$simulation, 1, which.max) - 1) / 2
multiplot(coords, sampleV)

system.time(
  V2 <- calcMultiPotential(W_SR = W_SR, W_LR = W_LR, distance.ref = seq(1, 10, 1),
     sample = sampleV, coords = coords, verbose = TRUE)
)

Wcat_LR <- W_LR
Wcat_LR@x <- findInterval(x = Wcat_LR@x, vec = distance.ref) - 1

system.time(
  V2 <- calcMultiPotential(W_SR = W_SR, W_LR = Wcat_LR, distance.ref = seq(1, 10, 1),
     sample = sampleV, coords = as.matrix(coords), verbose = TRUE, check.args = FALSE)
)
# quicker but arguments have to be correctly specified

multiplot(coords, V2)