statsmodels.genmod.families.family.NegativeBinomial.loglike¶

NegativeBinomial.loglike(endog, mu, var_weights=1.0, freq_weights=1.0, scale=1.0)¶

The log-likelihood function in terms of the fitted mean response.

Parameters:¶

endogndarray

Usually the endogenous response variable.

mundarray

Usually but not always the fitted mean response variable.

var_weightsarray_like

1d array of variance (analytic) weights. The default is 1.

freq_weightsarray_like

1d array of frequency weights. The default is 1.

scalefloat

The scale parameter. The default is 1.

Returns:¶

llfloat

The value of the loglikelihood evaluated at (endog, mu, var_weights, freq_weights, scale) as defined below.

Notes

Where \(ll_i\) is the by-observation log-likelihood:

\[ll = \sum(ll_i * freq\_weights_i)\]

ll_i is defined for each family. endog and mu are not restricted to endog and mu respectively. For instance, you could call both loglike(endog, endog) and loglike(endog, mu) to get the log-likelihood ratio.