Is it possible to use OpenBayes to learn by incrementally adding observations? At first glance, the inference engine's LearnMLParams(cases) appears to wipe out any previously learned distributions before parsing the new cases. Is this right?

Regards,
Chris

Yep i believe it works that way now. The learning is in uber beta mode, so all we were focussing on was the most basic case, though it isn't very difficult for us to add support for online learning, by just have the LearnMLParams rely on the existing CPDs. This isn't priority number 1 for us right now, but it is high on the list.

Unless you want to work on it... then it could be done real quick.

Elliot