Display patterns in orthogonal spatial and target coordinates

Usage

rotate_patterns(fit, spatial = "varimax", target = "varimax")

# S3 method for class 'pattern_view'
predict(object, ...)

Arguments

fit: A fitted pattern model or global result with a refit.
spatial, target: Orthogonal matrices, "varimax", or "none".
object: A pattern view.
...: Arguments passed to predict.pattern_fit.

Value

A pattern_view containing L_b = A Q_b, \(H = Q_b^T Q_t\), and L_t = C Q_t, in the fitted feature and whitened target coordinates. Thus \(L_b H L_t^T\) reconstructs \(A C^T\). Back-transforms, a basis identifier, and the Frobenius reconstruction error (evaluated without a full feature by target matrix) are retained.

Details

The underlying fit is unchanged. Predictions (including calibrated scores) and invariant maps delegate to that fit and are bit-identical. Display columns are not separately identified scientific components.