Excuse me, can put the LDA model and SVM together to make a new model? Because they two like measure is different, one is a bayesian probability, the other is the distance measure, the statistical model can combine the two contrary school