plan opなし → design(引数だけで決まる op —— 画像やデータの入力を取らない)import fullseye as fs; fs.ledger.doe_orthogonal_array(name='L9', factors=None, columns=None) -> 'dict' (実装を直接呼ぶなら import doe; doe.doe_orthogonal_array(name='L9', factors=None, columns=None) -> 'dict'、台帳から引くなら opsmultivar.get("doe_orthogonal_array"))Orthogonal array L4 / L8 / L9 / L12 / L16 / L18 / L27 from a published table -> design.
The arrays are copied from published tables (sources and the copy check in
:data:OA_SOURCES), not generated: Taguchi’s numbering of columns is what
shop-floor literature and the linear graphs refer to, and a generated array with
another column order would silently assign factors to interaction columns.
Published level symbols 1, 2 (, 3) map to coded −1, +1 (, or −1, 0, +1).
factors: int / names / (name, lo, hi[, unit[, scale]]) (default: one factor per
column). columns: 1-based column numbers to assign the factors to, in order
(default: the first k columns; for L18 column 1 is the two-level one).
The returned alias lists, for every pair of assigned factors, the columns their
interaction falls on and the share (see :func:_oa_interaction_share) — if such a
column also carries an assigned factor, that main effect is confounded with the
interaction. meta['strength2_ok'] re-checks orthogonality on the table actually
used, and meta['regular'] says whether interactions land on whole columns.
py -3.11 examples/doe_plan_tour.pypy -3.11 examples/multivar_inspection_tour.pypy -3.11 examples/poc_simcal_thin_film.pydesign を入力に取れる)doe_decode · rsm_fit · doe_main_effects · sens_morris · sens_sobol
plan)doe_lhs · doe_fractional_factorial · doe_central_composite · doe_box_behnken
Provenance: doe.py — MULTIVAR operator registry. この per-op ノートは tools/opdocs.py md が自動生成(手編集しない)。
© 2026 Kazufumi Furuse — Fullseye operator documentation. Licensed under Apache-2.0.