statsmodels.genmod.families.family.Tweedie

class statsmodels.genmod.families.family.Tweedie(link=None, var_power=1.0, eql=False)[source]

Tweedie family.

Parameters:

link : a link instance, optional

The default link for the Tweedie family is the log link. Available links are log and Power. See statsmodels.genmod.families.links for more information.

var_power : float, optional

The variance power. The default is 1.

eql : bool

If True, the Extended Quasi-Likelihood is used, else the likelihood is used (however the latter is not implemented). If eql is True, var_power must be between 1 and 2.

Notes

Logliklihood function not implemented because of the complexity of calculating an infinite series of summations. The variance power can be estimated using the estimate_tweedie_power function that is part of the statsmodels.genmod.generalized_linear_model.GLM class.

Attributes

Tweedie.link (a link instance) The link function of the Tweedie instance
Tweedie.variance (varfunc instance) variance is an instance of statsmodels.genmod.families.varfuncs.Power
Tweedie.var_power (float) The power of the variance function.

Methods

deviance(endog, mu[, var_weights, …]) The deviance function evaluated at (endog, mu, var_weights, freq_weights, scale) for the distribution.
fitted(lin_pred) Fitted values based on linear predictors lin_pred.
loglike(endog, mu[, var_weights, …]) The log-likelihood function in terms of the fitted mean response.
loglike_obs(endog, mu[, var_weights, scale]) The log-likelihood function for each observation in terms of the fitted mean response for the Tweedie distribution.
predict(mu) Linear predictors based on given mu values.
resid_anscombe(endog, mu[, var_weights, scale]) The Anscombe residuals
resid_dev(endog, mu[, var_weights, scale]) The deviance residuals
starting_mu(y) Starting value for mu in the IRLS algorithm.
variance
weights(mu) Weights for IRLS steps