Calibration

Calibration converts raw measures into set-membership scores. PyQCA provides crisp, piecewise fuzzy, and logistic fuzzy calibration utilities.

Crisp calibration

from qca import calibrate_crisp

calibrated = calibrate_crisp(
    data,
    value_col="raw_policy",
    threshold=5,
    out_col="policy",
    direction="high",
).df

Piecewise fuzzy calibration

from qca import calibrate_piecewise

calibrated = calibrate_piecewise(
    data,
    value_col="raw_capacity",
    out_col="capacity",
    anchors=(2, 5, 8),
).df

The anchors represent full non-membership, crossover, and full membership. When anchors are omitted, the utility can estimate them from quantiles.

Logistic fuzzy calibration

from qca import calibrate_logistic

calibrated = calibrate_logistic(
    data,
    value_col="raw_support",
    out_col="support",
    crossover=5,
).df

Calibration is a substantive analytical decision. Use Sensitivity analysis to inspect whether conclusions change across reasonable thresholds or fuzzy anchors.