I am trying to do some real time "text classification" using the paralell lda node. I want to train the lda, on a large document table (for example all recieved Whattsapp messages), to extract topics 1-n. Then reuse its model (distibution, words and weights) for a classification of new unseen messages. For my project it is mandatory to use the same 1-n topics built by using the same extracted words, and weights and so on.
Many of the other Cluster-, and Classification nodes provide some possibility to save this information as, model, PMML or simply clustercentres. So is there any way to accomplish this?
I would really love to get some Input from you