the maximum number of iterations in the EM algorithm. Alternatively,
maxit can be a list containing some or all parameters in the list
returned by
emCgm.control.
tolerance
convergence criterion for the EM algorithm. This may either be a
number or a vector of two numbers. By default, convergence is assumed
when the maximum relative change in the cell probability estimates is
less than
tolerance[1]. You may also specify
tolerance[2], in which
case convergence also occurs when the absolute change in the
log-likelihood from one iteration to the next is less than
tolerance[2]. By default, the log-likelihood is not used in
checking for convergence.
last
for the EM algorithm, the sequence of iteration numbers that are saved
is determined as the final
last iterations.
trace
if TRUE, additonal information is printed during the EM algorithm.
VALUE:
a list containing the control values described above.