Use the aa.gmm.predict() function on the output of the aa.gmm function to predict the cluster assignments for each [[ 11 12 1D], [ 21 ... D], [ D-1 DD]] point in a specified data set. Because GMM functions do soft assignments of data points to clusters (that is, GMM functions give probabilities that each data point is in each cluster), users can specify the top N most likely clusters for a given point and the probability that the point is a member of each of those clusters.