In order to quickly approximate maximum likelihood estimators from massive data, this study examines the optimal subsampling method under the A-optimality criterion (OSMAC) for generalized linear ...