Search found 7 matches

by heyifan
Tue Jan 11, 2022 2:24 am
Forum: GeNIe
Topic: learn parameters with EM
Replies: 3
Views: 1054

Re: learn parameters with EM

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 implem...
by heyifan
Sat Jan 08, 2022 8:19 am
Forum: GeNIe
Topic: learn parameters with EM
Replies: 3
Views: 1054

learn parameters with EM

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 b...
by heyifan
Tue Dec 21, 2021 2:46 am
Forum: GeNIe
Topic: learning parameter with missing values
Replies: 3
Views: 1248

Re: learning parameter with missing values

Hi HeyFan, Uniform distributions are as good as (or as bad as) a randomly chosen starting point and you should not view them as the best! To the contrary, it often leads to worse convergence. If you are worried about different convergence points, I suggest that you try a few randomizations and comp...
by heyifan
Sun Nov 28, 2021 7:09 am
Forum: GeNIe
Topic: learning parameter with missing values
Replies: 3
Views: 1248

learning parameter with missing values

I learnt that it is better to use 'Randomize' option when learning parameter with missing values. The problem is that using different random seeds leads to different learning results. How should I know which one is the optimized learning results? Maybe it is due to the size of training dataset is to...
by heyifan
Fri Apr 30, 2021 2:05 am
Forum: GeNIe
Topic: Learning Parameter algorithm
Replies: 5
Views: 2714

Re: Learning Parameter algorithm

The results should be valid, although a variable with just one state is just a constant, which carries no information. Constants are not allowed in structure learning but they will be accepted in parameter learning. The result of learning will be a large probability for the state represented in the...
by heyifan
Sun Apr 25, 2021 1:48 am
Forum: GeNIe
Topic: Learning Parameter algorithm
Replies: 5
Views: 2714

Re: Learning Parameter algorithm

Thanks! If I delete the node which only has one state, the CPT of other nodes learning by dataset won't be affected, right?
by heyifan
Tue Apr 13, 2021 3:32 am
Forum: GeNIe
Topic: Learning Parameter algorithm
Replies: 5
Views: 2714

Learning Parameter algorithm

I have constructed the structure of BN and have a dataset to learn the CPT. The dataset contains 90 records with states of 9 binary variables. The problem is that one of the nine varibles has only one state in all records. Will the parameter learning result still be valid? Thanks a lot if someone co...