MLSMU6 alternating least-squares unfolding

Multidimensional Least-Squares Metric Unfolding (Poole 1984; Bakker & Poole 2013). Alternates between respondent and stimulus coordinates, restarting from random seeds and keeping the lowest-stress fit.

Usage

morie_spatial_voting_mlsmu6(
  D,
  n_dims = 2L,
  max_iter = 200L,
  tol = 1e-06,
  n_restarts = 5L
)

Arguments

D

Respondent-by-stimulus distance/rating matrix.

n_dims

Number of latent dimensions.

max_iter

Maximum alternations per restart.

tol

Convergence tolerance on relative stress change.

n_restarts

Number of random restarts.

Value

A list with respondent_coords, stimulus_coords, stress, iterations, converged.

References

Poole, K. T. (1984). "Least Squares Metric, Unidimensional Unfolding." Psychometrika, 49(3). Bakker, R. and Poole, K. T. (2013).

Examples

set.seed(1)
D <- matrix(stats::runif(20 * 6), 20, 6)
morie_spatial_voting_mlsmu6(D, n_dims = 2, n_restarts = 1, max_iter = 50)
#> $respondent_coords
#>             [,1]        [,2]
#>  [1,]  0.9115420 -0.04226053
#>  [2,] -0.5826871 -1.16740152
#>  [3,]  0.1235497  0.07253461
#>  [4,]  0.3188333  1.15963197
#>  [5,]  0.1686051  1.67310829
#>  [6,] -0.2558838 -0.10035884
#>  [7,]  1.0246868  0.01455394
#>  [8,] -0.2282195 -1.15369288
#>  [9,]  1.4069451  0.57299100
#> [10,] -0.2022093 -0.29404778
#> [11,]  0.8543600  0.56864400
#> [12,]  1.6188930  0.75537690
#> [13,] -1.2192136  1.01123410
#> [14,] -0.3441043 -0.26472092
#> [15,] -0.2366259  0.58071442
#> [16,]  0.3633907 -1.11514651
#> [17,] -0.3653623 -0.40293543
#> [18,] -2.2032345 -0.44193450
#> [19,] -2.0298840 -1.66312389
#> [20,]  0.8766184  0.23683357
#> 
#> $stimulus_coords
#>             [,1]        [,2]
#> [1,]  0.16203645 -0.33191375
#> [2,] -0.06729308  0.51312240
#> [3,]  0.34329801 -0.18845019
#> [4,] -0.21968927  0.25653773
#> [5,] -0.39962810  0.09585039
#> [6,]  0.18127598 -0.34514657
#> 
#> $stress
#> [1] 111.1204
#> 
#> $iterations
#> [1] 50
#> 
#> $converged
#> [1] FALSE
#>