Usage
calcPotts_cpp(W_SR, W_LR, sample, rho, coords, site_order, iter_max, cv_criterion, test_regional, distance_ref, threshold, neutre, nbGroup_min, multiV,
last_vs_others, prior, type_reg, verbose)
Arguments
W_SR
the local neighborhood matrix. dgCMatrix. Should be normalized by row (i.e. rowSums(W_SR)=1).
W_LR
the regional neighborhood matrix. dgCMatrix. Should contain the distances between the observations (0 indicating infinite distance).
sample
the initial group probability membership. numeric vector.
rho
the value of the spatial regularisation parameters. numeric vector.
coords
the voxel coordinates. matrix.
site_order
a specific order to go all over the sites. integer vector.
iter_max
the maximum number of iterations. integer.
cv_criterion
the convergence criterion. numeric.
test_regional
should regional regularisation be used. logical.
distance_ref
the distance defining the several neighborhood orders relatively to W_LR. numeric vector.
threshold
the minimum value to consider non-negligible group membership. numeric.
neutre
the value to which the regional potential is set when the group is too small to compute it. numeric.
nbGroup_min
the minimum group size of the spatial groups required for performing regional regularization. integer.
multiV
should the regional potential range be specific to each spatial group ? logical.
last_vs_others
should the regional potential be computed for each class or only for the last versus the others ? logical.
type_reg
the type image to regularize : continuous or categorical ? logical.
prior
should the sample values be used as a prior ? logical.
verbose
should the running be be traced over iterations ? logical. 1 to display each iteration and 2 to display convergence diagnostics.