What is a BBN?
BBN stands for Bayesian Belief Net or Networks. It refers to graphical representations of random variables along with their dependencies. It is a powerful tool in testing various probabilities for a particular subject or relationship. Like in the field of medicine, when practitioners are presented with a variety of systems, the BBN method may be used to know the probable presence of different diseases.
In typical BBN graphs, there are arcs and nodes. The nodes are the graphical representations of the variables while the arc pertains to possible causes and relationship between the listed variables. In the basic example of determining the faults in a given product testing, variables may include possible causes that render the test inaccurate or invalid. These variables may include the following: very complex operations or testing, quality and capabilities of the testers, method and techniques used in the product testing, quality of supplies and equipment used, and many others. In this particular example, only one of the variables may be the actual cause, but this specific variable may be a result of another variable on the list. This is because certain variables may have dependencies or relationships with the others which need to be represented with arcs for proper interpretation.
The use of BBN is said to be valuable in determining “uncertain” concerns and issues. Using the example above, an observer may note that it is not necessary that all variables need to be present in order to have a fault in product testing. The graph merely states that some variable/s may be the actual cause of the testing fault and that there may be instances that some variables are dependent or happen as a result of its relationship with other variables.
BBN is typically used as model or guide in determining situations that may appear as vague, conflicting, or uncertain. It is said to be effective in situations that occur through cause and effect.