4

A DBN means that your model is replicated over each time discrete time step t, called slice. Variables of each slice can be connected together, as well a from previous slice to a latter slice (in this precise direction only). The probability tables and links remain the same in each slice if the model is stationary, which is the default case (otherwise they ...


2

They're not directly comparable. They do different things. They solve different problems. A Bayesian network is a probabilistic model of the relationship between multiple random variables. It is a generative model. It builds a model of the joint probability distribution between multiple random variables. It typically requires some priors or assumptions ...


2

You already have the result of the chain rule - this is your networks topology basically. Now you need to compute the joint. I would start from "root nodes", i.e. nodes with no parents, and proceed forward (I'm sure you can go the other way too, just more pain): Each of the "root nodes" has its "own" joint already. I.e. if your network was just one root ...


1

The maximum spanning tree is about summing weights; maximizing the probability is about multiplying weights. To convert multiplication to sums, take the logarithm. Hopefully that's enough for you to work out an algorithm for this task.


1

I'm realizing now that it's just regular matrix multiplication. For the case above, let $p(x_5|x_3)$ and $p(\bar{x_6}|x_2,x_5)$ be represented by $r\!\times\!s$ and $s\!\times\! t$ matrices, respectively, where $x_5$ can take on $s$ different values. Then, for each element $m_{ij}$ of $m_5$, we have that $$m_{ij}=\sum_{k=1}^s p_{ik}(x_5|x_3)p_{kj}(\bar{...


1

For anyone who shares this same query. I have found a link that answers it. Further, found a implementation as well.


Only top voted, non community-wiki answers of a minimum length are eligible