Calculate composite weights using PLS-PM
Calculate composite weights using the partial least squares path modeling (PLS-PM) algorithm Wold1975cSEM.
calculateWeightsPLS( .data = args_default()$.data, .S = args_default()$.S, .csem_model = args_default()$.csem_model, .conv_criterion = args_default()$.conv_criterion, .iter_max = args_default()$.iter_max, .PLS_ignore_structural_model = args_default()$.PLS_ignore_structural_model, .PLS_modes = args_default()$.PLS_modes, .PLS_weight_scheme_inner = args_default()$.PLS_weight_scheme_inner, .starting_values = args_default()$.starting_values, .tolerance = args_default()$.tolerance )
matrixof standardized or unstandardized data (indicators/items/manifest variables). Possible column types or classes of the data provided are: "
double" or "
ordered" and/or "
character" (converted to factor), or a mix of several types.
The (K x K) empirical indicator correlation matrix.
A (possibly incomplete) cSEMModel-list.
Character string. The criterion to use for the convergence check. One of: "diff_absolute", "diff_squared", or "diff_relative". Defaults to "diff_absolute".
Integer. The maximum number of iterations allowed. If
iter_max = 1and
.approach_weights = "PLS-PM"one-step weights are returned. If the algorithm exceeds the specified number, weights of iteration step
.iter_max - 1will be returned with a warning. Defaults to
Logical. Should the structural model be ignored when calculating the inner weights of the PLS-PM algorithm? Defaults to
FALSE. Ignored if
.approach_weightsis not PLS-PM.
Either a named list specifying the mode that should be used for each construct in the form
"construct_name" = mode, a single character string giving the mode that should be used for all constructs, or
NULL. Possible choices for
modeare: "modeA", "modeB", "modeBNNLS", "unit", "PCA", a single integer or a vector of fixed weights of the same length as there are indicators for the construct given by
"construct_name". If only a single number is provided this is identical to using unit weights, as weights are rescaled such that the related composite has unit variance. Defaults to
NULLthe appropriate mode according to the type of construct used is chosen. Ignored if
.approach_weightis not PLS-PM.
Character string. The inner weighting scheme used by PLS-PM. One of: "centroid", "factorial", or "path". Defaults to "path". Ignored if
.approach_weightis not PLS-PM.
A named list of vectors where the list names are the construct names whose indicator weights the user wishes to set. The vectors must be named vectors of
"indicator_name" = valuepairs, where
valueis the (scaled or unscaled) starting weight. Defaults to
Double. The tolerance criterion for convergence. Defaults to
A named list. J stands for the number of constructs and K for the number of indicators.
A (J x K) matrix of estimated weights.
A (J x J) matrix of inner weights.
A named vector of modes used for the outer estimation.
The convergence status.
TRUEif the algorithm has converged and
FALSEotherwise. If one-step weights are used via
.iter_max = 1or a non-iterative procedure was used, the convergence status is set to
The number of iterations required.