simcal_uncertainty — SIMCAL check op

• データ種: calresult → calresult

• 呼び出し: import fullseye as fs; fs.ledger.simcal_uncertainty(result=None, method='laplace', n_boot=40, level=0.95, seed=0) -> 'dict' (実装を直接呼ぶなら import simcalcheck; simcalcheck.simcal_uncertainty(result=None, method='laplace', n_boot=40, level=0.95, seed=0) -> 'dict'、台帳から引くなら opssimcal.get("simcal_uncertainty"))

使い方

Standard deviations and intervals of the determined knobs: Laplace (inverse Fisher) or parametric bootstrap.

Only knobs that kept a point estimate are covered; knobs without one are held at the end point (the intervals are conditional on them — stated in note). laplace: covariance = inverse of J^T W J at the estimate on the unit-scaled knobs, intervals u ± z sd mapped back through each knob's scale (a log knob gets an asymmetric interval), std by the delta method. bootstrap: n_boot refits to the fitted prediction plus Gaussian noise of the noise model's variance (Efron 1979), percentile intervals at level. A result with no point estimate is refused. Under a model discrepancy both are optimistic (simcal_discrepancy says so).

詳しい使い方ガイド

• sim_to_real_calibration ファミリ ガイド

参考(サンプルデータ・文献)

• サンプルデータ カタログ(DL URL / ライセンス) — 2-D は skimage.data(BSD/public)+ 合成、3-D は実データ源(Stanford/PDS 等)の DL URL。

• 演算子の来歴・参考文献 — この op 族の元になった研究/手法の出典。

• アルゴリズムの正典(著者・年)と用途は上記ファミリ使い方ガイドに記載。

実行できる例(この op を実際に呼ぶ検証済みサンプル)

• poc_simcal_thin_film — py -3.11 examples/poc_simcal_thin_film.py

型が繋がる次の op(calresult を入力に取れる)

simcal_identifiability · simcal_validate · simcal_discrepancy · simcal_report

同カテゴリ(check)

simcal_identifiability · simcal_validate · simcal_discrepancy


*Provenance: simcalcheck.py — SIMCAL 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.