Search found 430 matches

by marek [BayesFusion]
Tue Mar 12, 2024 6:39 pm
Forum: GeNIe
Topic: Strange behavior of TruncNormal
Replies: 7
Views: 363

Re: Strange behavior of TruncNormal

In GeNIe, you can see all samples generated on the left hand side in the Value tab. GeNIe displays the number of samples generated for any bar in the histogram when you hover over that bar. Distribution visualizer has the same functionality. We will extend the description in the manual to account fo...
by marek [BayesFusion]
Sat Mar 02, 2024 9:44 pm
Forum: GeNIe
Topic: Strange behavior of TruncNormal
Replies: 7
Views: 363

Re: Strange behavior of TruncNormal

You are proposing an interesting solution that will work well in many case, certainly in this case. Let us look at the simplified model with just three variables. We do get funny (in terms of being large) number of samples in forward inference for small and large values of evidence in v1 (e.g., 0..1...
by marek [BayesFusion]
Tue Feb 27, 2024 8:21 pm
Forum: GeNIe
Topic: Strange behavior of TruncNormal
Replies: 7
Views: 363

Re: Strange behavior of TruncNormal

I have looked at your model and I believe that this is a theoretical problem that is solvable at the modeling stage. Please note that the domain of v2 is {0,10} and when you have a combination of {0,1} at v1 and {-2,-1} at error, you generate a value for v2 that is out of bounds. This is handled eas...
by marek [BayesFusion]
Fri Dec 15, 2023 3:38 pm
Forum: GeNIe
Topic: Coding boolean statements
Replies: 2
Views: 630

Re: Coding boolean statements

Hi Fabrice, What you want to do mixes values (high, moderate, and low) with uncertainty about them (probability of "low" > 0.7). Bayesian networks are models consisting of variables, each of which has its domain (these are values), and interactions among them, which can be best thought of ...
by marek [BayesFusion]
Fri Nov 24, 2023 9:42 pm
Forum: GeNIe
Topic: enter evidence, virtual evidence, and controlling values
Replies: 4
Views: 898

Re: enter evidence, virtual evidence, and controlling values

Hi Yan, Yes, you are correct in that if you want to see the effect of both sport and drinking water, you should control both. In the example that you gave, the effect on health of controlling both will be the same as the effect of observing both. In fact, you can see the effect in the CPT for node C...
by marek [BayesFusion]
Thu Nov 23, 2023 2:37 pm
Forum: GeNIe
Topic: enter evidence, virtual evidence, and controlling values
Replies: 4
Views: 898

Re: enter evidence, virtual evidence, and controlling values

Hi Yan, Let me try to answer your questions. 1. My understanding is that entering evidence can only give 100% for one state of node A, while virtual evidence can give the probability distribution (e.g., 20%, 40%, 40%, assuming there are three states) for all states of node A. Is it right? If so, doe...
by marek [BayesFusion]
Tue Oct 31, 2023 9:37 pm
Forum: GeNIe
Topic: max sensitivity Genie 2.4
Replies: 4
Views: 1896

Re: max sensitivity Genie 2.4

Hi Camilla, I'd love to look at this model to make sure that I am understanding your question correctly. Perhaps we can meet online? Please shoot me a private message so that we can set the time. Alternatively, perhaps you can share with me the relevant part of the model? You can use GeNIe obfuscati...
by marek [BayesFusion]
Tue Oct 10, 2023 4:30 pm
Forum: GeNIe
Topic: Log-transformation data
Replies: 2
Views: 731

Re: Log-transformation data

You can't do it in GeNIe. What you can do is transform your data in Excel by creating for a variable x a new variable/column named, e.g., log_x) and then save the file as .CSV and load it into GeNIe. You can use log_x instead of x in learning.
I hope this helps,

Marek
by marek [BayesFusion]
Fri Sep 15, 2023 8:46 pm
Forum: GeNIe
Topic: structure learning
Replies: 4
Views: 1195

Re: structure learning

Hi Yan, I have looked at your data and have managed to replicate the problem. Clearly, there is a bug in PC and, as you wrote, the algorithm ignores the max. adjacency size parameter. We will work on it and will fix it in the next release. When I face a problem like yours, I usually play with the al...
by marek [BayesFusion]
Thu Sep 14, 2023 1:02 pm
Forum: GeNIe
Topic: structure learning
Replies: 4
Views: 1195

Re: structure learning

Hi Yan, Let me try to answer your questions. (1) This is hard to say, as the two algorithms work on different principles. I would try both and look for commonalities in the output. This especially that your data set is rather small. Please play with the significance level alpha in PC -- it has a big...
by marek [BayesFusion]
Fri Sep 08, 2023 12:47 pm
Forum: GeNIe
Topic: Combining expert elicitation and emprical data
Replies: 2
Views: 1006

Re: Combining expert elicitation and emprical data

Absolutely! This is in fact what we have been advising when the data set is small (a "small data set" is, of course, a relative term and depends on the complexity of the network but we are dealing with too small data sets most of the time :-)). You can start with a rough estimation of the ...
by marek [BayesFusion]
Fri Jun 09, 2023 7:28 pm
Forum: GeNIe
Topic: parameter learning and validation in GeNIe 4.0 vs 3.0
Replies: 3
Views: 1543

Re: parameter learning and validation in GeNIe 4.0 vs 3.0

I believe that case counting should be at least as good as the EM algorithm. It is a special case of EM.
I hope this helps,

Marek
by marek [BayesFusion]
Fri Mar 24, 2023 12:09 pm
Forum: GeNIe
Topic: Validation using data with known diagnoses
Replies: 11
Views: 3233

Re: Validation using data with known diagnoses

Correct, if there are missing values in that column -- then other variable (or the rest of the network in general) will impact the learned probabilities. Also correct if there are few records (like 10 :-)), as EM starts with a prior in each distribution. The data just change this prior. If there are...
by marek [BayesFusion]
Thu Mar 23, 2023 8:40 pm
Forum: GeNIe
Topic: Validation using data with known diagnoses
Replies: 11
Views: 3233

Re: Validation using data with known diagnoses

Parameter learning is essentially learning probability distributions from data. If all values were present and the feature would be present in two cases out of ten, the probability distribution learned would be 0.2/0.8. However, when two of the 10 values are absent, I cannot predict without seeing t...
by marek [BayesFusion]
Thu Mar 23, 2023 6:12 pm
Forum: GeNIe
Topic: Validation using data with known diagnoses
Replies: 11
Views: 3233

Re: Validation using data with known diagnoses

Parameter learning is based on the EM algorithm, which is an iterative procedure that replaces the missing values with the most likely values and then does the counting, like you did. I am not surprised that the parameters that you obtained are slightly different from the parameters obtained by EM. ...