sensitivity op• Data kinds: design × signal → table
• Call: import fullseye as fs; fs.ledger.sens_morris(design=None, y=None, negligible=0.05) -> 'dict' (to call the implementation directly, import sensitivity; sensitivity.sens_morris(design=None, y=None, negligible=0.05) -> 'dict'; from the registry, opsmultivar.get("sens_morris"))
Morris elementary effects -> mu, mu*, sigma per factor (Morris 1991; mu* Campolongo et al. 2007).
design must come from doe_morris (its trajectory bookkeeping is used) and y holds the model output for its rows in order. For trajectory t and step s that moves factor i by ±delta: EE_i = (y_s - y_{s-1}) / (±delta) in unit-cube units, and also per natural unit of the factor (ee_natural; on a log-scale factor the step is in log10 units).
mu (mean EE) can cancel when effects change sign; mu_star (mean |EE|) does not and is the ranking measure; sigma (standard deviation, ddof = 1) is large when the effect depends on where the other factors are (non-linear or interacting). class per factor: negligible (mu* below negligible x the largest mu*), linear/additive (sigma < 0.1 mu*: the effect is the same everywhere — a convention of this op, stated), otherwise non-linear or interacting.
• design_of_experiments family guide
• Sample-data catalog (download URLs / licences) — 2-D uses skimage.data (BSD/public domain) plus synthetic images; 3-D lists download URLs for real data sources (Stanford, PDS, …).
• Operator provenance and references — the sources of the research/methods this op family came from.
• The canonical algorithm (author, year) and its uses are named in the family usage guide above.
• sensitivity_tour — py -3.11 examples/sensitivity_tour.py
table as input)—
sensitivity)doe_morris · doe_saltelli · sens_sobol
*Provenance: sensitivity.py — MULTIVAR operator registry. This per-op note is generated by tools/opdocs.py md (do not hand-edit).*
© 2026 Kazufumi Furuse — Fullseye operator documentation. Licensed under Apache-2.0.