genmod.families.family.Gaussian()

statsmodels.genmod.families.family.Gaussian

class statsmodels.genmod.families.family.Gaussian(link=) [source]

Gaussian exponential family distribution.

Parameters:

link : a link instance, optional

The default link for the Gaussian family is the identity link. Available links are log, identity, and inverse. See statsmodels.family.links for more information.

Attributes

Gaussian.link a link instance The link function of the Gaussian instance
Gaussian.variance varfunc instance variance is an instance of statsmodels.family.varfuncs.constant

Methods

deviance(endog, mu[, scale]) Gaussian deviance function
fitted(lin_pred) Fitted values based on linear predictors lin_pred.
loglike(endog, mu[, scale]) Loglikelihood function for Gaussian exponential family distribution.
predict(mu) Linear predictors based on given mu values.
resid_anscombe(endog, mu) The Anscombe residuals for the Gaussian exponential family distribution
resid_dev(endog, mu[, scale]) Gaussian deviance residuals
starting_mu(y) Starting value for mu in the IRLS algorithm.
weights(mu) Weights for IRLS steps
doc_statsmodels
2017-01-18 16:09:20
Comments
Leave a Comment

Please login to continue.