Generalized Means Rotation

Computes a rotation axis that best represents a target variable's contribution to the ENA point space, after controlling for covariates via Lasso (coordinate-descent, k-fold CV). Mirrors rENA::ena.rotate.by.generalized().

Usage

generalized_means_rotation(
  V,
  x_model_matrix,
  x_target,
  x1_cols,
  x_categorical,
  x_n_groups,
  x_subset,
  has_y,
  y_model_matrix,
  y_target,
  y1_cols,
  y_categorical,
  y_n_groups,
  n_lambda = 50L,
  k_folds = 5L,
  lasso_eps = 0.01
)

Arguments

V

Numeric matrix (n_units x n_connections) — ENA points.

x_model_matrix

Numeric matrix (n_units x p) — model matrix for x axis.

x_target

Numeric vector length n — target variable (raw float for numeric; 0-based integer codes for categorical).

x1_cols

Integer vector of 0-based column indices in x_model_matrix that belong to the target variable (unpenalized).

x_categorical

Logical — TRUE if target is categorical.

x_n_groups

Integer — number of distinct groups (categorical only).

x_subset

Integer vector of 0-based row indices to subset for the x-axis GMR step (pass integer(0) to use all rows).

has_y

Logical — TRUE to compute a second GMR axis.

y_model_matrix

Numeric matrix (n_units x p) — model matrix for y axis (ignored when has_y = FALSE).

y_target

Numeric vector — y-axis target (ignored when has_y = FALSE).

y1_cols

0-based column indices for y target (ignored when has_y = FALSE).

y_categorical

Logical (ignored when has_y = FALSE).

y_n_groups

Integer (ignored when has_y = FALSE).

n_lambda

Length of the Lasso lambda path (default 50).

k_folds

Cross-validation folds for lambda selection (default 5).

lasso_eps

lambda_min = lasso_eps * lambda_max (default 0.01).

Value

List with rotation, eigenvalues, column_names.

Details

The caller is responsible for building x_model_matrix (e.g. via model.matrix()) and identifying x1_cols (0-based indices of the target variable's columns, which receive penalty_factor = 0). Categorical targets should be encoded as 0-based integers.