Storage of learning for CPT

[loiclegoff]

Hi!
I work about a prediction project.
Your development is very attractive. For my project, I have a lot of paths for a learning (about 100/200 M).
I am newbie to BIG DATA and I don't know how I can store my tree after learning step.
At the moment, I binarize my tree and different tables of CPT (LT, II, ..).
Have you got a solution about this problem? How did you store your tree ?
Did you ever store you tree as a graph database?
Thanks!

[tedgueniche]

Because this is a common problem, we include an example on how to serialize a trained model into a file and how to load the model from file. This can be very useful to train the model offline.

SerializePredictor.java: Shows how to train a model and save it into a file. The model can then be archived or transferred to another machine. It also shows how to load the model back in memory.