statsmodels.discrete.discrete_model.Logit.loglikeobs¶

Logit.loglikeobs(params)[source]¶

Log-likelihood of logit model for each observation.

Parameters:

params : array-like

The parameters of the logit model.

Returns:

loglike : ndarray (nobs,)

The log likelihood for each observation of the model evaluated at params. See Notes

Notes

$\ln L=\sum_{i}\ln\Lambda\left(q_{i}x_{i}^{\prime}\beta\right)$

for observations $i=1,...,n$

where $q=2y-1$ . This simplification comes from the fact that the logistic distribution is symmetric.