On September 8, 2008 - we have released new versions of the HUGIN Graphical User Interface (v7.0) and HUGIN Decision Engine (v7.0).
The main new features of this release are:
Limited memory influence diagrams (LIMIDs) replace influence diagrams as the tool for modeling decision problems. The LIMID representation removes two fundamental assumptions of the traditional influence diagram: A non-forgetting decision maker and a total order of the decisions. By removing these two assumptions, the LIMID representation implies a significant improvement in the support for modeling decision problems. In addition to supporting the LIMID representation, the tool has been updated to supply the user with additional information related to a decision model (e.g., show policies and both decision and chance nodes have a probability distribution and an expected utility function determined by the current strategy).
The new Expression Builder Wizard makes it easier to build an expression by dividing the process into steps that allow the user to concentrate on building one simple expression at a time while keeping an overview of the complete expression.
HUGIN Graphical User Interface v7.0:
The HUGIN Graphical User Interface has been improved with various new features. This includes:
The new constraints panel of the Parameter Sensitivity Analysis (PSA) Wizard allows the user to specify constraints on the posterior probability of a hypothesis variable and apply the smallest change necessary to satisfy it. The changes can be confirmed in the PSA panel.
The color map in the PSA panel displays Min/Max parameter sensitivity values instead of parameter values.
Improved performance of the parameter sensitivity analysis in the PSA Wizard.
The hypothesis variable drop-down box in the PSA panel marks nodes with evidence as the analysis cannot be performed on instantiated nodes.
The clear button in the PSA panel now also initializes the network.
The graph info table in the PSA panel supports the selection of a line description (row) that results in displaying the selected line in the graph. (Useful when performing analysis on all states of the hypothesis.)
The graph (PSA panel) displays upper and lower bounds for the sensitivity function.
Display of Markov Blanket.
The Data Analyzer Tool in the Data Preprocessing panel of the Learning Wizard now allows the discretization of multiple numerical variables.
The Learning Wizard and the EM Learning Wizard have been improved with functionality for reporting different pieces of information relevant for the learning process.
ROC analysis is now available as a part of the Analysis Wizard
The results displayed in the “Cases/Belief”-tab of the Analysis Wizard can be exported as a CSV file.
A Correlation Analysis Dialog allows the user to investigate the correlation between any pair of discrete chance nodes in a Bayesian network.
The Min-Max panel of the Evidence Sensitivity Dialog shows the results as graphs and numbers instead of as tables of numbers.
Better support for running the HUGIN Graphical User Interface on Windows Vista.
Other minor improvements.
Finally, efforts have been put into improving the stability of the HUGIN Graphical User Interface.
HUGIN Decision Engine v7.0:¶
The HUGIN Decision Engine has been extended with the following features:
HUGIN API libraries for the Windows platforms are now provided for Visual Studio 2008 (in addition to Visual Studio 6.0, Visual Studio .NET 2003, and Visual Studio 2005).
The “h_domain_compile” function now initializes the junction trees using the available evidence (if any).
The “h_class_create_domain” function now assigns unique “meaningful” names to the nodes of the created domains. This is accomplished by allowing “dotted names” for nodes in domains (but not for nodes in classes). This change affects the lexical syntax of node names in files (NET, data, case, and node list files) and in strings containing expressions.
Sensitivity analysis: It is now possible to compute sensitivity data for multiple output probabilities simultaneously. This allows for easy solution of constraints on the output probabilities.
- Limited memory influence diagrams (LIMIDs) replace influence diagrams as the tool for modeling decision problems. This has several implications:
There is no “no-forgetting” assumption in a LIMID: All information links must be explicitly represented in the network.
A total order of the decisions is no longer required.
Evidence specified for a LIMID must satisfy the “free will” condition: A chance node can’t be observed before all decisions in the ancestral set of the chance node have been made. Moreover, only simple instantiations are now permitted as evidence for decision nodes.
Computing optimal decisions is now a separate operation (that is, they are no longer computed as part of inference).
Decision nodes now have tables. These tables represent decision policies and can be specified in the same way (including the use of models) as conditional probability and utility tables. However, they are usually computed by the “h_domain_update_policies” operation.
In LIMIDs, there are no constraints on elimination orders used for triangulation.
The overall expected utility of the decision problem is provided by the new “h_domain_get_expected_utility” function.
“h_node_get_belief” and “h_node_get_expected_utility” now accept (discrete) chance as well as decision nodes as arguments, and “h_node_get_expected_utility” also accepts utility nodes.
The usage conditions and the treatment of information links has changed for d-separation analyses in LIMIDs.
The “nodes” argument of “h_domain_get_marginal” may now contain decision as well as chance nodes. Also, decision nodes are no longer required to be instantiated and propagated.
The “h_domain_simulate” operation no longer requires decision nodes to be instantiated and (for compiled domains) propagated.
The EM algorithm and the “h_domain_get_log_likelihood”, “h_domain_get_AIC”, and “h_domain_get_BIC” functions no longer require decision nodes to be instantiated in all cases of the data set. Instead, the evidence in a given case is only required to be “valid”.
Sensitivity analysis: Sensitivity functions where the input parameter is associated with a decision node can now be computed. Also, decision nodes are no longer required to be instantiated and propagated.
The arguments of “h_node_get_entropy” and “h_node_get_mutual_information” may now be decision as well as chance nodes. Also, decision nodes are no longer required to be instantiated and propagated.
The “h_domain_save_to_memory” operation is disabled for LIMIDs.
Calculation of conflict of evidence is not supported for LIMIDs.
The NET format has been extended in order to represent policies.
The HKB format has changed.