Search found 392 matches

by marek [BayesFusion]
Thu Jun 30, 2022 9:09 pm
Forum: GeNIe
Topic: Too many states in nature nodes
Replies: 3
Views: 28

Re: Too many states in nature nodes

I roughly see what you want to accomplish. I'm afraid your CPT in node PorductHierarchyID will have 40,000 conditional probability distributions and given that the node itself has 50 states, it will have 2,000,000 probabilities. If you multiply these by 8 bytes needed to represent a floating point n...
by marek [BayesFusion]
Tue Jun 28, 2022 7:16 pm
Forum: GeNIe
Topic: Too many states in nature nodes
Replies: 3
Views: 28

Re: Too many states in nature nodes

Generally, probabilistic representations are wonderfully efficient but only if there are independencies in the domain. Still, probability tables grow exponentially with the number of parents. The base of the exponent is the number of states in the node and parents. Having a node with thousands of st...
by marek [BayesFusion]
Tue Jun 21, 2022 7:15 pm
Forum: GeNIe
Topic: randomize parameter
Replies: 2
Views: 55

Re: randomize parameter

Hi Yan, I believe you are asking about randomization in the context of the EM algorithm (please correct me if I am wrong). In validation, this parameter also refers to the EM algorithm part (learning the parameters at each cross-validation step). It is an empirical finding that randomization gets yo...
by marek [BayesFusion]
Thu Jun 16, 2022 11:39 am
Forum: GeNIe
Topic: Help with influence diagram
Replies: 4
Views: 50

Re: Help with influence diagram

Clearly, your influence diagram is too complex. Please keep in mind that probabilistic inference is worst case NP-hard and it is not that difficult to construct models that are not computable. Quite likely if a model cannot be solved by SMILE/GeNIe, you cannot solve it with any software on Earth. Th...
by marek [BayesFusion]
Tue Jun 14, 2022 3:08 pm
Forum: GeNIe
Topic: Help with influence diagram
Replies: 4
Views: 50

Re: Help with influence diagram

Your model is quite impressive (in terms of size and complexity). I am not sure you can solve it using our default ID algorithm, as it will involve calculating the expected utilities over 2^19 (524,288) different policies. This is, of course, when you change back the RCM* nodes to decision nodes. Ar...
by marek [BayesFusion]
Thu Jun 09, 2022 6:51 pm
Forum: GeNIe
Topic: query about noisyMAX algorithm
Replies: 9
Views: 254

Re: query about noisyMAX algorithm

I see. I misunderstood your example but have have clarified it sufficiently. When A and B have an unobserved common parent C, you can write the following: P(A,B)=Sigma_over_i P(A|B,C_i)P(B|C_i)P(C_i) You write this formula for all possible states of C_i. Because both A and B are dependent on C, you ...
by marek [BayesFusion]
Wed Jun 08, 2022 9:33 am
Forum: GeNIe
Topic: query about noisyMAX algorithm
Replies: 9
Views: 254

Re: query about noisyMAX algorithm

Please look at a good Bayesian networks textbook. Pearl 1988 perhaps? When two adjacent variables are dependent, you can see it in the CPT: the distributions are different for different states of the parent. P(A,B)=P(A|B)P(B) for A and B being dependent. When they are independent, P(A|B)=P(A), i.e.,...
by marek [BayesFusion]
Mon Jun 06, 2022 1:51 pm
Forum: GeNIe
Topic: query about noisyMAX algorithm
Replies: 9
Views: 254

Re: query about noisyMAX algorithm

Don't the formulas in my earlier post help? P(E)=P(E|AB)P(AB)+P(E|A~B)P(A~B)+P(E|~AB)P(~AB)+P(E|~A~B)P(~A~B) vs. P(E)=P(E|AB)P(A)P(B)+P(E|A~B)P(A)P(~B)+P(E|~AB)P(~A)P(B)+P(E|~A~B)P(~A)P(~B) You can construct a simple model in GeNIe and plug in some numbers in the CPTs to see whether the theoretical ...
by marek [BayesFusion]
Fri Jun 03, 2022 7:51 pm
Forum: SMILE
Topic: the probability method in java
Replies: 1
Views: 138

Re: the probability method in java

What exactly do you mean by a fuzzy Bayesian network? There is a good chance that GeNIe does not implement it. We implement standard Bayesian networks, hybrid Bayesian networks (a mixture of discrete and continuous variables, very close to systems of simultaneous structural equations), dynamic Bayes...
by marek [BayesFusion]
Fri Jun 03, 2022 7:48 pm
Forum: SMILE
Topic: Calculate function
Replies: 1
Views: 61

Re: Calculate function

What exactly do you mean by a Fuzzy RB?
Cheers,

Marek
by marek [BayesFusion]
Fri Jun 03, 2022 7:47 pm
Forum: GeNIe
Topic: diagnostic values are too small
Replies: 1
Views: 32

Re: diagnostic values are too small

The meaning of the diagnostic values is purely theoretical and there are no units. They are essentially cross-entropies between a node and the fault in question. When it is 0.006, I agree that it is rather small. The interpretation is that when you learn the value of this node, you will not learn mu...
by marek [BayesFusion]
Sat Apr 30, 2022 11:17 am
Forum: GeNIe
Topic: How to set evidence to a binomial equation node by external counters of successes
Replies: 4
Views: 338

Re: How to set evidence to a binomial equation node by external counters of successes

Please excuse me for a delay in answering -- I was traveling and had a lot to catch up after the travel. Thank you for the clear explanation of the problem. Now I see what you are after. I think the attached model will do the trick. You will need to enter evidence for the nodes Observation? (either ...
by marek [BayesFusion]
Thu Apr 21, 2022 11:47 am
Forum: GeNIe
Topic: network structure is too complex
Replies: 1
Views: 63

Re: network structure is too complex

Hi Bahman, I have looked at your network and it is not really Node13 that is the culprit. Please note that your network will not update even if you remove this node. Generally, it is not that easy to predict the size of the junction tree in advance without trying triangulation. Your network is quite...
by marek [BayesFusion]
Wed Apr 20, 2022 5:52 pm
Forum: SMILE
Topic: ERROR
Replies: 5
Views: 403

Re: ERROR

I’m on the road, so I will be very brief. A small percentage of invalid samples is generally not a problem. Gray arcs tell you that there is something wrong with the definition of the child. They mean, essentially, that they are not necessary (the child is independent of the parent), which means tha...
by marek [BayesFusion]
Wed Apr 20, 2022 7:50 am
Forum: SMILE
Topic: ERROR
Replies: 5
Views: 403

Re: ERROR

Thank you for the network. I can see that the domain of the node Blood_Sugar is [0.4,4.4] and the definition is Blood_Sugar=Weibull(451.545,0.365594). The samples from your Weibull span between practically zero and almost 400K. Many, many samples fall outside of the interval [0.4,4.4] and this is wh...