FAQ

1) What is CSV format? Comma Seperated values is a text data format with each value in a row of data delimited with a comm(,). It is available from the Save As dialogue on spreadsheet programs such as Microsoft Excel and is frequently used as an human-readable data interchange format. It can also be easily created using text editors.
e.g.
Node A, Node B, Node C
26,12,34
35,23,48...etc
 

2) What is Structure learning? This is where the program learns the structure of a bayesian belief network from a data-file (which contains a set of cases i.e. observed data).
Hugin can use a variety of statistical algorithms which analyse the conditional independence of nodes in the network. They have differing strengths and weaknesses e.g. speed, completeness or level of expert human intervention.
Algorithms available include:
PC algorithm - see http://download.hugin.com/webdocs/manuals/7.3/Htmlhelp/descr_PC_algorithm_pane.html
NPC algorithm - see http://download.hugin.com/webdocs/manuals/7.3/Htmlhelp/descr_NPC_algorithm_pane.html
Greedy search-and-score algorithm
Chow-Liu tree algorithm - see http://en.wikipedia.org/wiki/Chow-Liu_tree
Rebane-Pearl polytree algoritm -see http://en.wikipedia.org/wiki/Polytree
Tree Augmented Naive Bayes algorithm.

3) What is a Domain model? A Doman model is a mathematical model of a problem domain. On this web-site a Domain model will refer to a Bayesian network (or an influence which, in principle, is a Bayesian network augmented with decision facilities such as utility functions, decision variables and information orderings).