learn parameters with EM

The front end.
Post Reply
heyifan
Posts: 7
Joined: Tue Aug 25, 2020 2:05 pm

learn parameters with EM

Post by heyifan »

The structure of the network has been defined and conditional probability table have been elicited from an expert. I want to use the new data set as an additional source of information over the existing network. How does 'Keep original' option works? From my point of view, if I set 'Confidence' to be 100, then a data file containing 100 records will be generated based on original parameters, and new parameters will be learnt with the generated data file combining new data set. Is that right?
marek [BayesFusion]
Site Admin
Posts: 350
Joined: Tue Dec 11, 2007 4:24 pm

Re: learn parameters with EM

Post by marek [BayesFusion] »

Hi Heyifan,

What you wrote is essentially correct, although the mechanics of this is somewhat more sophisticated than reducing the existing parameters to 100 records. If we did that, there would be a chance of never having records for some combination of values of parent nodes. The way it is implemented, if there are no records corresponding to a combination of values of parent nodes, the original value of the parameters will be preserved.

I hope this helps,

Marek
heyifan
Posts: 7
Joined: Tue Aug 25, 2020 2:05 pm

Re: learn parameters with EM

Post by heyifan »

marek [BayesFusion] wrote: Mon Jan 10, 2022 3:35 pm Hi Heyifan,

What you wrote is essentially correct, although the mechanics of this is somewhat more sophisticated than reducing the existing parameters to 100 records. If we did that, there would be a chance of never having records for some combination of values of parent nodes. The way it is implemented, if there are no records corresponding to a combination of values of parent nodes, the original value of the parameters will be preserved.

I hope this helps,

Marek
Thank you, Marek. Really appreciate your help. And 'Keep original' option works more like Bayesian estimation which makes use of a priori knowledge, rather than MLE, is that right?
marek [BayesFusion]
Site Admin
Posts: 350
Joined: Tue Dec 11, 2007 4:24 pm

Re: learn parameters with EM

Post by marek [BayesFusion] »

Yes, you could view it this way.
Cheers,

Marek
Post Reply