-
this method is used in estimation of parameters like mean,
variance etc. of a probability distribution function
-
also this method is used when direct access to data is not
possible or if some data is missing
-
EM consists of two steps: an Expectation step followed
by a Maximization step
-
the expectation step is with respect to unknown variables using
the current estimate of parameters. At the start of the process
the starting values for parameters are assumed
-
the maximization step then provides a new estimate of
the parameters
-
the two steps are iterated till convergence is reached