BayesFusion Support Forum

Hi Nour, I'm afraid I am not aware of any tool that can do that. Parameter (and equation?) learning in a hybrid system is a complex problem and it depends very much on the form of the equation, distributions, etc. No general solution. When we move into that direction, which we definitely will, we wi...

Hi Nira, I'm afraid there is no good answer to your question, as the answer depends on the domain knowledge and what you want to accomplish. Imagine somebody is asking you: What do I put in the Excel cells? What equations and what numbers? These questions are similar to what you have asked. The answ...

Hi Bahman, Sorry for the delay. As I suggested in my previous answer, we support our software but not really modeling effort, which requires domain knowledge and is always time consuming. We can do this for you on a consulting basis if this is something that would interest you. Let me just answer yo...

Hi Nira, I don't see any problems with it except for that it is in a very initial stage of construction. The grey arcs tell you that you need to look at the definition in the child nodes (I explained it in another query and it is also explained in GeNIe manual, section "Structural analysis"...

Hi Nira, A discrete node can have continuous parents. In that case, the discretization that you specify in the Discretization tab will be used to creating the CPT for the child node. Please look at the section "Hybrid models" in the chapter "Equation-based and hybrid models" in G...

Hi Nour, SMILE has everything that GeNIe has, so it certainly implements hybrid networks. Grey color denotes arcs that are not necessary in the sense of having no impact on the child variable. Whenever you create a new arc, it will have grey color because the CPT in the node contains (identical) uni...

Hi Nira,

Yes, this is correct. You will need to discretize the continuous variables in your data set (please make sure that they match the states in the network!). SMILE's implementation of the EM algorithm will work for mixture of discrete and continuous variables.

I hope this helps,

Marek

HI Paolo, I'm not 100% sure if I understand your question completely but I have modified your model (added a third set of variables) and created an additional node through which you can enter evidence. I'm attaching this model. Does this do what you wanted? If not, which I suspect is the case, what ...

Hi Bahman, You are correct -- I used the CPT created by automatic discretization for copying and pasting into the discrete nodes into the DBN. You can do the same for any dynamic model -- you just need to reflect on the correct structure, i.e., what influences what. You can get this information from...

The problem is quite simple and quite independent of whether the node Effect is a NoisyOR or is just a regular chance variable. P(E) is equal to: 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) In the first case, i.e., without the Ancestor, A and B are independent of each other and ...

Hi Yan, You are pretty much right. The paper of Coupe and van der Gaag looks at this problem precisely the way we do. GeNIe calculates the derivative of the posterior probability in question taken over each individual parameter. The algorithm GeNIe's implementation is based on is described in Kjærul...

Hi Bahman, Generally, a continuous (equation) variable can contain any function from the list of functions supported by GeNIe (please look at the manual or the tree of functions on the right-hand side in node definition. Conditional functions include If(), Choose() and Switch(). I believe that you w...

Hi Bahman, When I look at the table of the variables in this model, I am also getting confused: Something is funny here, even though it does seem to be a DBN. We are quite surely dealing with N different temporal models (different devices?) that don't interact with each other outside of the initial ...

Hi Bahman, Most definitely, you can edit the unrolled network any way you please. You can also edit the equations. We do not support equations in DBNs yet but it is a matter of time -- we just haven't had time to do that but we will most certainly return to this issue. Essentially, we need to extend...

Hi Bahman, You should be able to create a network like the one in the picture using SMILE. It may be a challenging network in terms of the number of parameters and perhaps inference but well, SMILE is very, very fast, so you may not even notice the time that it takes to update this network. If the s...