Search found 309 matches

by marek [BayesFusion]
Thu Dec 03, 2020 12:30 pm
Forum: SMILE
Topic: SMILE Error in EM.learn
Replies: 9
Views: 545

Re: SMILE Error in EM.learn

I see. Parameter learning does not pose the theoretical questions that I described in my previous post. You can easily learn individual CPTs from your data set. Please do be cautious when learning these from different data sets, as this may pose theoretical problems. A paper that I co-authored with ...
by marek [BayesFusion]
Thu Dec 03, 2020 9:37 am
Forum: SMILE
Topic: SMILE Error in EM.learn
Replies: 9
Views: 545

Re: SMILE Error in EM.learn

Hi EAly, It sounds like a good way to proceed if you have really many variables, although only an approximate procedure. I assume you mean learning the structure from subsets of columns (variables). The risk of doing this is that in each subset you may leave out some variables that are possibly invo...
by marek [BayesFusion]
Wed Dec 02, 2020 11:08 pm
Forum: SMILE
Topic: SMILE Error in EM.learn
Replies: 9
Views: 545

Re: SMILE Error in EM.learn

Hi Ealy, Your model is too large and too complex to even load at the moment, so you will need to simplify it. I understand that you have generated it automatically through SMILE, so you haven't lost anything. You need to realize that the complexity of a Bayesian network model grows with the complexi...
by marek [BayesFusion]
Mon Nov 23, 2020 7:23 pm
Forum: GeNIe
Topic: Learning in a hybrid network with discretized data
Replies: 3
Views: 814

Re: Learning in a hybrid network with discretized data

Hi, I have looked at your network and the data set. GeNIe will not allow you to learn parameters in hybrid networks. If you want to derive a usable network from this, you will need to discretize it first (Network menu/Discretize). This will create a discrete network that will accept the data from th...
by marek [BayesFusion]
Wed Nov 04, 2020 9:20 am
Forum: SMILE
Topic: How log likelihood value is calculated in EM?
Replies: 12
Views: 2356

Re: How log likelihood value is calculated in EM?

GeNIe/SMILE do not switch to approximate algorithms automatically -- it is the user's responsibility to designate the default inference algorithm. If exact algorithms fail, I recommend the EPIS sampling algorithm, which is quite likely the state of the art algorithm for approximate inference in disc...
by marek [BayesFusion]
Tue Oct 27, 2020 12:18 pm
Forum: GeNIe
Topic: Two equations for one node
Replies: 3
Views: 428

Re: Two equations for one node

Of course! You can make the logical conditions as complex as you wish with "Logical functions" such as And(), Max(), Min(), Or() and Xor(). You can also arbitrarily nest If() statements.
Cheers,

Marek
by marek [BayesFusion]
Mon Oct 26, 2020 2:58 pm
Forum: GeNIe
Topic: Two equations for one node
Replies: 3
Views: 428

Re: Two equations for one node

Sure, this is possible, although you would need two separate nodes (this is closer to what your problem really is, I believe). Please look at the model in our model repository (should be also included among the example models on your disk) called "Heat Equations Autodiscretized Hybrid.xdsl". The nod...
by marek [BayesFusion]
Sun Oct 25, 2020 5:30 pm
Forum: GeNIe
Topic: Model Comparison through log(p)
Replies: 4
Views: 432

Re: Model Comparison through log(p)

Not at all. log(p) is the probability of the whole data set given the model and it depends strongly on the data set size and the number of variables. Generally, the probability of a combination of values of nodes within a record is very low. Please verify this by choosing a record in your data set, ...
by marek [BayesFusion]
Sun Oct 25, 2020 11:03 am
Forum: GeNIe
Topic: Dynamic Bayesian Network (DBN) for time series
Replies: 8
Views: 740

Re: Dynamic Bayesian Network (DBN) for time series

Here are two models that you may want to explore. The "Foxes Rabbits Equations.xdsl" model is a hybrid model that contains the difference equations describing the dynamic model. I used this model to derive the CPTs for the Dynamic Bayesian Network in "Foxes Rabbits DBN 20.xdsl". I discretized the co...
by marek [BayesFusion]
Sun Oct 25, 2020 10:36 am
Forum: GeNIe
Topic: Cross validation in GeNIe with different proportions
Replies: 1
Views: 238

Re: Cross validation in GeNIe with different proportions

Sure but you have to split the file into the two parts (70% and 30%) yourself. Then learn the model from the 70% part and perform "Test only" on the 30%. k-fold cross-validation and leave-one-out are much more efficient, although the training phase in these is parameter learning only and not the str...
by marek [BayesFusion]
Sat Oct 24, 2020 5:56 pm
Forum: GeNIe
Topic: Dynamic Bayesian Network (DBN) for time series
Replies: 8
Views: 740

Re: Dynamic Bayesian Network (DBN) for time series

Hi Teng Wang, It is actually not that bad -- I have created DBN models in GeNIe that were based on systems of difference equations, such as predator-prey models. Unlikely you will find another tool -- GeNIe has the most flexible implementation of hybrid Bayesian networks. Finally, we will extend DBN...
by marek [BayesFusion]
Sat Oct 24, 2020 11:06 am
Forum: GeNIe
Topic: Dynamic Bayesian Network (DBN) for time series
Replies: 8
Views: 740

Re: Dynamic Bayesian Network (DBN) for time series

Hi Teng Wang, It is the first -- we have not married equations and DBNs but you are right that it is possible in theory. One way that I deal with it now is creating static networks with equations and deriving CPTs for the equation nodes through discretization. Once the nodes are discretized, I copy ...
by marek [BayesFusion]
Fri Oct 23, 2020 12:08 pm
Forum: GeNIe
Topic: Counting of relationship between a specific variable and the main parent
Replies: 5
Views: 513

Re: Counting of relationship between a specific variable and the main parent

Actually, there is but it is a trick that has not really been designed for this purpose. It will work for discrete Bayesian networks. When the network is open, please invoke the "Strength of Influence" mode (either through a choice in the Network menu or by pressing a corresponding button in the too...
by marek [BayesFusion]
Thu Oct 22, 2020 4:53 pm
Forum: GeNIe
Topic: Model Comparison through log(p)
Replies: 4
Views: 432

Re: Model Comparison through log(p)

Hi Hakan, Log(p), ranging from minus infinity to zero, is not model accuracy but rather an expression of fit of the model to the data. Its numerical value is best used in comparisons between multiple runs of the learning algorithm -- the higher the number, the better (please keep in mind that the nu...
by marek [BayesFusion]
Thu Oct 22, 2020 12:42 pm
Forum: SMILE
Topic: Unexpected evaluation result
Replies: 8
Views: 993

Re: Unexpected evaluation result

Hi Sverre, If the nodes a1 and b1 are not evidence nodes, you can marginalize them to simplify your model. This should work fine. Please just remember that marginalization removes the two nodes while preserving the properties of the rest of the model. After you have marginalized these nodes, entered...