SMACOF for Individual Differences

Variant of smacof::smacofIndDiff() but with regularization (weightlam) in case of singularity in weight matrices (e.g., due to missingness) and optional weight scaling (weights) which multiples weights for each data source by some amount.

Usage

smacofIndDiff(
  delta,
  ndim = 2,
  type = c("ratio", "interval", "ordinal", "mspline"),
  constraint = c("indscal", "idioscal", "identity"),
  weights = NULL,
  weightmat = NULL,
  weightlam = 0,
  init = "torgerson",
  ties = "primary",
  verbose = FALSE,
  modulus = 1,
  itmax = 100,
  eps = 1e-06,
  spline.degree = 2,
  spline.intKnots = 2
)

Arguments

delta

A list of dissimilarity matrices or a list objects of class dist

ndim

Number of dimensions

type

MDS type: "interval", "ratio", "ordinal" (nonmetric MDS), or "mspline"

constraint

Either "indscal", "idioscal", or "identity" (see details)

weights

Vector of length equal to the number of distance matrices. If provided, weights for each distance matrix are multiplied by the corresponding value in weights. Default is NULL which will give each distance matrix the same weight (i.e., scaled by 1).

weightmat

Optional matrix with dissimilarity weights

weightlam

Regularization parameter for weight matrices to avoid singularity matrices. Default is 0 (i.e., no regularization). If regularization is needed, set to a small value (e.g., 1e-6).

init

Matrix with starting values for configurations (optional)

ties

Tie specification for non-metric MDS

verbose

If TRUE, intermediate stress is printed out

modulus

Number of smacof iterations per monotone regression call

itmax

Maximum number of iterations

eps

Convergence criterion

spline.degree

Degree of the spline for "mspline" MDS type

spline.intKnots

Number of interior knots of the spline for "mspline" MDS type