How does one fit an 8-variable finite mixture model in SAS? I.e., assuming k segments, the model to be fitted is that the joint density of scores X=(X1,...,X8) we observe, f(X), can be written as f(X) = α1f(X/1) + α2f(X/2) + … + αkf(X/k). Here the α’s are the proportions of customers in each segment and f(X/i) is the joint density of the scores X conditional on belonging to segment i, i=1,…,k. The mixing proportions (α’s) and conditional parameters for each f(X/i) have to be estimated from the data. We could probably assume an 8-variate normal distribution for f(X) with a pre-specified inter-variable correlation structure. Further, how does one obtain information criteria for evaluating the number of component densities k? And lastly, how can one take into account sampling weights in fitting this model?