mt op• 데이터 종류: signal → table
• 호출: import fullseye as fs; fs.ledger.spc_mt_sn_ratio(md)(구현을 직접 호출하려면 import spc; spc.spc_mt_sn_ratio(md), 원장에서 가져오려면 opsspc.get("spc_mt_sn_ratio"))
> 이 연산자의 설명은 아직 번역이 없습니다. 원문을 그대로 싣습니다.
Larger-the-better SN ratio of abnormal-sample distances, in decibels.
MT-method item selection asks which features earn their place: run the unit
space with a subset of features, score the *abnormal* samples, and keep the
subset whose distances are largest. The figure of merit is the
larger-the-better signal-to-noise ratio
eta = -10 log10( (1/n) sum_i 1 / MD_i^2 )
an inverse-square average, so one abnormal sample that the subset fails to
separate drags the whole score down — which is the behaviour wanted when the
cost of a miss dominates. The caller drives the subsets (classically by an
orthogonal array); this operator is the closed-form score for one subset.
`md` is a 1-D array of distances for the abnormal samples. Returns a dict
with `sn_ratio_db, the harmonic_mean_md_sq it came from, and n`.
For a constant `MD == d the sum collapses and eta == 20 log10(d)`
exactly — pinned in the tests, and the reason a subset that puts every
abnormal sample at distance 10 scores exactly 20 dB.
Raises `ValueError`: an empty or non-1-D *md*, a non-positive entry (a
zero distance would divide by zero: an abnormal sample sitting exactly on the
unit-space centre means the subset separates nothing), or non-finite input.
• 샘플 데이터 카탈로그(DL URL / 라이선스) —— 2-D 는 skimage.data(BSD/public)+ 합성, 3-D 는 실데이터 소스(Stanford/PDS 등)의 DL URL.
• 연산자의 내력·참고문헌 —— 이 연산자 족의 바탕이 된 연구/기법의 출처.
• 알고리즘의 정전(저자·연도)과 용도는 위의 패밀리 사용 가이드에 적혀 있습니다.
• poc_spc — py -3.11 examples/poc_spc.py
table 를 입력으로 받는 것)msa_anova_table · msa_gauge_rr · msa_bias_linearity · msa_attribute_agreement · gum_standard_uncertainty · gum_propagate · gum_expanded · gum_monte_carlo
mt)spc_mt_unit_space · spc_mt_distance
*Provenance: spc.py — SPC 연산자 레지스트리. 이 op 노트는 tools/opdocs.py md 가 자동 생성합니다(직접 편집하지 마세요).*
© 2026 Kazufumi Furuse — Fullseye operator documentation. Licensed under Apache-2.0.