henrikbostrom/crepes
A Python library for conformal prediction that extends scikit-learn classifiers and regressors with calibrated uncertainty estimates.
Collecting fresh signals — velocity needs a few days of history.
star history
The crepes package provides implementations of standard and Mondrian conformal classifiers, regressors, and predictive systems. It works on top of any standard scikit-learn classifier or regressor, transforming their predictions into well-calibrated p-values and cumulative distribution functions, or prediction sets and intervals with coverage guarantees. The package includes tools for computing non-conformity scores and difficulty estimates.