plan op• 数据种类:无 → design(仅由参数决定的算子 —— 不接受图像或数据输入)
• 调用: import fullseye as fs; fs.ledger.doe_fractional_factorial(factors=5, p=2, generators=None, center=0) -> 'dict'(要直接调用实现,import doe; doe.doe_fractional_factorial(factors=5, p=2, generators=None, center=0) -> 'dict';从台账取用则 opsmultivar.get("doe_fractional_factorial"))
由生成元构造的二水平正则部分因子设计 2^(k-p)（Box 与 Hunter 1961）-> design。
> 以下的详细说明为原文 —— 摘要与标题已翻译。
The first k - p factors (letters A, B, C ... skipping I) form a full factorial in standard (Yates) order, A changing fastest. Each generator "D=ABC" sets an added factor's column to the product of base columns ("E=-ABD" for the other half). generators=None takes the recommended generators for (k, p) from the NIST/SEMATECH e-Handbook table (FF_NIST, 5.3.3.4.7 Table 3.17, after Box, Hunter & Hunter); p=0 or generators=() gives the full factorial 2^k.
The defining relation (every product of generator words I = ABCD = ...), the resolution (length of the shortest word: III = main effects aliased with two-factor interactions, IV = main effects clear but 2fi aliased with each other, V = 2fi clear) and the alias table of main effects and two-factor interactions are returned in meta and alias. They are computed from the generators, and the returned matrix is checked against them (each aliased pair of columns is identical up to sign).
center adds centre runs (coded 0) for a curvature check; they do not change the alias structure.
• 示例数据目录(下载 URL / 许可证) —— 2-D 用 skimage.data(BSD/公有领域)加合成图,3-D 给出真实数据源(Stanford/PDS 等)的下载 URL。
• 算子来历与参考文献 —— 该算子族所依据的研究/方法出处。
• 算法的正典(作者、年份)与用途见上面的族使用指南。
• doe_plan_tour — py -3.11 examples/doe_plan_tour.py
design 作为输入)doe_decode · rsm_fit · doe_main_effects · sens_morris · sens_sobol
plan)doe_lhs · doe_orthogonal_array · doe_central_composite · doe_box_behnken
*Provenance: doe.py — MULTIVAR 算子登记表。本条目由 tools/opdocs.py md 自动生成(请勿手工编辑)。*
© 2026 Kazufumi Furuse — Fullseye operator documentation. Licensed under Apache-2.0.