Normalize and optimize a directed (ordered) network accumulation

Thin wrapper around model that applies the ONA-specific defaults: SVD rotation, zero-network exclusion from centering, and center-to-origin post-processing.

Usage

model(
  x,
  ...,
  rotate.using = "svd",
  rotation.params = NULL,
  rotation.set = NULL,
  center_to_origin = TRUE
)

Arguments

x

An ena.ordered.set produced by accumulate(ordered = TRUE) or tma::accumulate_contexts.

...

Additional arguments forwarded to rENA::model.

rotate.using

Character. "svd" (default) for standard principal components rotation; "mean" for means rotation.

rotation.params

Additional parameters for the rotation function (passed as rotate_params to rENA::model). Required with rotate.using = "mean": list(list(a, b)), the two groups whose means define the first dimension, as logical vectors over the units or as unit names.

rotation.set

Optional rotation to reuse instead of fitting one: a fitted ena.ordered.set (its $rotation is used), that $rotation list, or an rENA ENARotationSet. The data are normalised and centred as usual, projected with the supplied rotation matrix, and the supplied node positions are kept rather than re-optimised, so new units can be placed in an existing model's space.

center_to_origin

Logical. When TRUE (default) the projected space is translated so the mean of all points is at the origin.

Value

An ena.ordered.set with calculated points and node positions.