Evaluating a Bayesian network

A Bayesian network is a graph of probabilistic dependencies. Nodes in the graph are events, and edges represent conditional dependence. We can build a network from prior knowledge to find out new probabilistic properties of the events.

We will use Haskell's probabilistic functional programming library to evaluate such a network and find interesting probabilities.

Getting ready

Install the probability library using cabal as follows:

$ cabal install probability

We will be representing the following network. Internalize the following figure to get an intuitive grasp of the variable names:

Event C depends on events A and B. Meanwhile, events D and E depend on event C. Through the power of the Probabilistic Functional Programming library, in this recipe, we will find the probability of event E given only information about event D.

How to do it…

1. Import the following packages:

   import qualified Numeric.Probability.Distribution as Dist
   import Numeric.Probability.Distribution...