Cluster Analysis - EM tab

Select the EM tab of the Cluster Analysis dialog box to access options to specify parameters for the EM clustering algorithm. See also the Introductory Overview for details.

Random seed

Specify the random number generator seed to be used when initializing the classification probabilities (weights).

Minimum increase of log-likelihood

Specify a minimum increase in the log-likelihood value over consecutive iterations; when the log-likelihood from one iteration to the next is smaller than this value, the iterative clustering algorithm terminates, and the Results dialog box is displayed.

Distributions for continuous variables

Specify the distribution for each continuous variable; by default, all continuous variables are treated as following the Normal distribution. After clicking this button, a variable selection dialog box for three variables lists is displayed, where you can assign the (previously selected) continuous variables to one of three distributions: normal, log-normal, or Poisson.