Get an instance of GLMInfluence with influence and outlier measures
If true, then observed hessian is used in the hat matrix computation. If false, then the expected hessian is used. In the case of a canonical link function both are the same.
The instance has methods to calculate the main influence and outlier measures as attributes.