From: xiao.xiong.1981@gmail.com   
      
   I have a general question to ask:   
      
   There are a set of supervised training data which belong to two classes   
   A and B. Each training sample is a multi-dimensional vector and the   
   covariance matrix of the features are not diagonal, that is the feature   
   elements are correlated. Assume the underlying pdf is completely known.   
   We can use two Gaussian mixture models (GMM) for the two classes   
   separately and build a Bayesian classifer.   
      
   My quesetion is:   
      
   Is it possible to use some kinds of feature transformation method, such   
   as Linear Discriminative Analysis, to project the features into a new   
   space for better discriminative power?   
      
   [ comp.ai is moderated ... your article may take a while to appear. ]   
      
   --- SoupGate-Win32 v1.05   
    * Origin: you cannot sedate... all the things you hate (1:229/2)