To calculate the expected value of the missing data given the observed data and current parameter estimates.

To update the parameter estimates by maximizing the expected log-likelihood found in the E-step.

To compute the variance of the observed data.

To determine the number of iterations required for convergence.