SMACOF stress minimisation

Iterative majorisation algorithm for metric MDS.

Usage

morie_spatial_voting_smacof(
  D,
  n_dims = 2L,
  max_iter = 300L,
  tol = 1e-06,
  weights = NULL,
  init = NULL
)

Arguments

D

Symmetric dissimilarity matrix.

n_dims

Number of dimensions.

max_iter

Maximum iterations.

tol

Convergence tolerance on stress change.

weights

Optional weight matrix (defaults to uniform).

init

Optional initial configuration (n x n_dims).

Value

A list with coordinates, stress, iterations, converged.

References

De Leeuw, J. (1977). "Applications of Convex Analysis to Multidimensional Scaling." In Recent Developments in Statistics, 133-145. Borg & Groenen (2005).

Examples

D <- as.matrix(dist(matrix(rnorm(40), 10)))
morie_spatial_voting_smacof(D)
#> $coordinates
#>             [,1]       [,2]
#>  [1,]  1.7546107  0.5994812
#>  [2,] -0.1151257  2.7746612
#>  [3,] -0.2282911 -0.3829059
#>  [4,]  0.7247420  0.5720256
#>  [5,] -1.3321778  1.6768343
#>  [6,] -1.6843927 -1.7005330
#>  [7,]  1.0249369 -1.0000806
#>  [8,]  0.1357115 -1.4742940
#>  [9,]  0.6002775 -1.7562888
#> [10,] -0.8802913  0.6911000
#> 
#> $stress
#> [1] 11.60122
#> 
#> $iterations
#> [1] 98
#> 
#> $converged
#> [1] TRUE
#>