The nodes are one of the basics in constructing a HUGIN network (Bayesian network, LIMID, or object-oriented network). Currently, HUGIN has five types of nodes that differ in appearance and semantics.

  • Discrete Chance Nodes - represent random variables with a discrete finite state space.

  • Continuous Chance Nodes - represent random variables with a continuous infinite satet space having a Gaussian (normal) distribution.

  • Decision Nodes - represent decisions or actions with a discrete state space.

  • Utility Nodes - represent the utility function one or more decisions are to be compared on.

  • Instance Nodes - represent instances of other networks, enabling object-oriented specification of networks (both Bayesian networks and influence diagrams).

  • Function Nodes - computes a value based on expressions

The graph of a network consist of the nodes and the links.