convergence threshold. When
abs(new.deviance-old.deviance)/(old.deviance + epsilon) < epsilon
the algorithm is assumed to have converged.
maxit
maximum number of iterations
trace
should iteration details be printed while
glm() is fitting the model.
VALUE:
a list is returned, consisting of the three parameters, conveniently
packaged to supply the
control argument to
glm(). The values for
glm.control()
can be supplied directly in a call to
glm(). These
values are then filtered through
glm.control() inside
glm().
SEE ALSO:
,
,
.
EXAMPLES:
glm(formula, family, control = glm.control(maxit = 15))
glm(formula, family, maxit = 15) # these are equivalent