Keyboard shortcuts

Press ← or → to navigate between chapters

Press S or / to search in the book

Press ? to show this help

Press Esc to hide this help

name: op-attr-factor-score
description: Per-row score on one common factor of a vector of numeric members, from an inline fit with the factor-analysis matrix operator's knobs, loadings, order and sign; regression or Bartlett scores; fit_on fits on a subset and still scores every row.
kind: operator
category: ATTR
operator: ATTR_FACTOR_SCORE
type: reference
applies_to: process, compose, predict
examples_tags: [multivariate, feature-engineering, streaming-friendly]

Row-level scalars; no Response.Components.

Use when

Adds to every row its score on one factor behind several related numeric fields: an estimate of the hidden trait they measure together.

Questions it answers:

  • What is each respondent’s score on the satisfaction factor behind these rating questions?
  • Can I score every row on the two constructs a factor analysis found, to compare segments on them?

Params

  • factor — required, 1-based. method — regression (default) or bartlett. missing — listwise only.
  • factors, extraction, rotation, kappa, gamma — the factor fit’s knobs (MAT_FACTOR’s); no repair.
  • fit_on — "filtered" (default) or a filterer array picking the FIT rows; every filtered row is scored.
  • weight — fit only; weight-0 rows are scored, not fitted.

Slot: exactly one of vector (a vectors entry) / fields; no field.

Inputs

Members: u4…u64, f32/f64 source columns.

Output

One f64 per record: zᵀw on fit-standardized members (psych factor.scores); mean 0 over the fit rows; null member → null.

Reading the output

  • value: The row’s estimated score on the chosen factor, on the scale of the fields standardized over the fitted rows: 0 is an average row, positive is toward the end where the fields loading positively on the factor are high.
    • Caveat: The direction is MAT_FACTOR’s convention (each factor’s loadings sum positive), not a meaning: check its loadings on the same fields before calling a high score good or bad.
    • Caveat: After an oblique rotation (promax, oblimin) the factors correlate, so scores on two factors move together.

Gotchas

  • Fit is cohort-wide, never per group; groupers or features force buffered. Two-pass, never merged.
  • All null + warning: PULSE_MATRIX_SINGULAR, _INSUFFICIENT_N, _NOT_IDENTIFIED, _HEYWOOD (Bartlett). Scored + warning: a regression Heywood case, PULSE_MATRIX_NOT_CONVERGED.

See