# statsmodels.genmod.families.family.Family¶

class `statsmodels.genmod.families.family.``Family`(link, variance)[source]

The parent class for one-parameter exponential families.

Parameters: link : a link function instance Link is the linear transformation function. See the individual families for available links. variance : a variance function Measures the variance as a function of the mean probabilities. See the individual families for the default variance function.

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