Skip to content

Investigate when BetaUnivariate raises an error and how to fix it #455

Description

@fealho

Error Description

It is possible for the BetaUnivariate to raise an error when fitting to some data:

from copulas.univariate import BetaUnivariate
BetaUnivariate().fit(np.array([0.7122827 , 0.04830956, 0.54410219, 0.04173127, 0.54462469, 0.54565197, 0.05497849, 0.07792652, 0.6817948 , 0.19735519]))

# Fails with the following:
scipy.stats._warnings_errors.FitError: Optimization converged to parameters that are outside the range allowed by the distribution.

The error happens when fitting the univariate in the following line:

a, b, loc, scale = beta.fit(X, loc=loc, scale=scale)

Expected Behavior

Investigate what causes scipy to fail and what can be done to improve the convergence rate.

Additional Context

This happens because "scipy.stats.fit doesn't guarantee success".

Note: On the rare occasion when this error happens, the synthesizers - either in Copulas or in SDV - would still work, since they have a fallback distribution for such cases.

Activity

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Metadata

Metadata

Assignees

No one assigned

    Labels

    Type

    No type

    Projects

    No projects

      Milestone

      No milestone

      Relationships

      None yet

      Development

      No branches or pull requests

      Issue actions