[ANN] GFIlogisticRegression

Hello,

Announcing my third Julia package: Generalized fiducial inference for logistic regression.

Roughly speaking, fiducial statistics is “Bayesian statistics without prior”: similarly to the Bayesian inference, the fiducial inference is based on a distribution which describes the uncertainty on the model parameters; but it does not require a prior distribution. Generalized fiducial inference has good frequentist properties, e.g. a fiducial 95%-confidence interval is almost a frequentist 95%-confidence interval.

Cheers.

2 Likes