HUGIN is an easy to use app for building and running Bayesian networks. You can build new and update existing models by adding or deleting nodes, states and links between nodes as well as specifying the conditional probability distribution for each node using a nice UI. The UI implements different gestures to perform certain actions on the model.
You can insert and propagate evidence using the traditional HUGIN message passing algorithm. This ensures fast and correct belief update in the light of evidence. The posterior distribution of each node given evidence is computed automatically without an explicit compilation step.
The app supports Bayesian networks with discrete chance nodes specified in the HUGIN knowledge base (hkb) format. This means that Bayesian networks can be shared with other users as well as opened in the HUGIN Graphical User Interface and Decision Engine. This makes the app a great supplement to these tools.
The HUGIN app is a great tool for learning about Bayesian networks. It can be used by teachers and students interested in the technology that is applied to solve many real-world tasks involving reasoning under uncertainty in artificial intelligence and machine learning.
The app can handle models with up to forty nodes each having as many as eight states. Tables can be as large as 128 parameters with a total size of 2048 parameters.