The Observation Theory EncyclopediaFrom TSKAboutBy kindBy chapterBy Lean fileLedgerProvenance

effective rank concept

DefinitionThe number of directions a matrix really uses, the square of the sum of its eigenvalues over the sum of their squares, defined for a nonnegative spectrum that is not all zero, with no cutoff to choose. Equation 0.7. Also participation ratio.
ExampleEigenvalues (1, 1, 1, 1) have effective rank 4, and (1, 0.12, 0.06, 0.02) have effective rank about 1.4.
BookData Mining as Observation, draft 0.2, commit f3914f0; entry id effective-rank, kind concept.
Statusmeasures [demonstrated]. Corrections: none recorded.
Defining equation

Book equation 0.7.

Assumptions and scope
  • The effective rank is the square of the sum of the eigenvalues over the sum of their squares, between one and the dimension for a nonnegative spectrum, with no cutoff to choose.
  • It is not invariant under a change of basis of the input, unlike the trace pairing and the rank, and the ledger row that says so is the one that licenses comparing read operators across coordinate systems.
  • An allocation over a spectrum with effective rank below two concentrates on one direction, and the allocation report says so before giving a gain.
  • It is defined for a nonnegative spectrum that is not all zero, and it lies between one and the dimension.
Prior artnone recorded
Evidencegeometric-observation/claims/LEDGER.md:30, readscope/readscope/spectrum.py:35-70, turboquant-pro/turboquant_pro/read_allocation.py:244-307, gtc-prototype/docs/SPECTRUM_FINDINGS.md:10-18, lean/DataMiningAsObservation/EffectiveRank.lean
Reviewedsemantic review 2026-09-06; generated 2026-09-10 from records at the commits on the provenance page.
directioneigenvalue
The number of directions a spectrum really uses, with no cutoff to choose.

Equation

Book equation 0.7.

\[r_{\mathrm{eff}}=\frac{\big(\sum_i\lambda_i\big)^{2}}{\sum_i\lambda_i^{2}}.\]

Conditions

Conditions are curated in entries.toml rather than read from a record.

Ledger

First stated

The participation ratio of a spectrum, applied to read operators in readscope, readscope/readscope/spectrum.py:35-70, and chapter 0 section 0.4 of Data Mining as Observation.

Measurements

Where the book states it Numbers, as the book’s sources table records them Source
chapter 4 section 4.2 effective rank as participation ratio, energy rank readscope/readscope/spectrum.py:35-70
chapter 4 section 4.2 allocation report, gain over uniform, concentration caution below effective rank 2 turboquant-pro/turboquant_pro/read_allocation.py:244-307
chapter 9 section 9.4 explained ratios, effective rank 5.19 of 8, convergence with a second method, property of the representation not the space gtc-prototype/docs/SPECTRUM_FINDINGS.md:10-18

Failures and corrections

none

Invariance envelope

none declared

Machine checked

lean/DataMiningAsObservation/EffectiveRank.lean, theorems sq_sum_le, effRank_le, sum_sq_le_sq_sum, one_le_effRank, effRank_const, effRank_single, at observation-data-mining f3914f0; what the check covers is stated in the book’s appendix C.

Used in

Data Mining as Observation primer L, chapters 0, 4, 9, 11, 14.

Related

read operator; read subspace; water-filling; recognizer.

See also

Book equations stated beside the entry’s terms, not defining it: 4.2.

Sources-table rows that share a record with the entry without naming it: chapter 14 section 14.3, chapter 14 section 14.5, chapter 14 section 14.7.

Status

Generated 2026-09-10 by encyclopedia/generate.py; book at observation-data-mining f3914f0; the commit of every record is listed in the encyclopedia’s provenance.

← early stoppingeigenvalue, eigenvector →