Today we are releasing a new version of the HUGIN software (v7.8). The main new feature of this release is the introduction of “explanation” facilities in the HUGIN Decision Engine.
The new “explanation” facilities of the HUGIN Decision Engine makes it more efficient and straightforward for the application developer to include explanation features in her own implementations. Explanation is defined as the task of identifying the impact of subsets of the evidence on the probability of an hypothesis.
HUGIN Graphical User Interface v7.8
The HUGIN Graphical User Interface has been improved with various new features. These include:
- The explanation functionality of the Evidence Sensitivity Dialog is now based on API functionality of the HUGIN Decision Engine.
- A Show Decision Utilities Dialog has been added to the HUGIN GUI. This allows the user to compute and see a table of the utilities for a specific decision from which the policy is identified using Single Policy Updating. This functionality can be used to inspect the table of utilities producing a policy for a decision.
- The Table View has been extended with a CG Table Wizard for Continuous Gaussian (CG) nodes. The CG Table Wizard allows the user to define mathematical expressions for the calculation of the mean and variance of a CG node. This functionality is only available in the HUGIN GUI.
- The policy table of a decision node can be copied and pasted to another decision node.
- The Value of Information Dialog now supports the calculation of a normalized mutual information measure between all subsets of nodes.
- The Input Panel of the HUGIN GUI Run-Time mode has been extended with functionality to collapse parts of the panel. This is useful when the panel includes a lot of information, for instance, a high number of levels in an object-oriented network.
- The Data Frame window has been extended with functionality to compute the ROC and confusion matrix for classification models. The ROC and confusion matrix are useful measures for determining the performance of a classifier.
- Two new parameters have been introduced for the total-weight triangulation method. These parameters can be used to produce better triangulations as well as speed up the triangulation process
- An “initial triangulation” can be specified: If a good triangulation is already known, then this can be used to speed up the search for a better triangulation.
- The search for minimal separators can be sped up by discarding separators larger than a prespecified size. This also makes it possible to handle larger prime components (without having to split them into smaller graphs).
- Functionality to reorder the states of a labelled discrete node has been included. This function also updates tables according to the new order of states.
- Other minor improvements.
Finally, efforts have been made into improving the performance of the HUGIN Graphical User Interface.
HUGIN Decision Engine v7.8
The HUGIN Decision Engine has been extended with the following features:
- A set of functions implementing explanation features. This enables the user to identify the impact of subsets of the evidence on the probability of a hypothesis using a single HUGIN function as well as functions to access the scores and subsets.
- A function to reorder the states of a decision node has been included. This function also updates tables according to the new order of states.