libDAI
Public Member Functions | Public Attributes | List of all members
dai::JTree::Properties Struct Reference

Parameters for JTree. More...

#include <dai/jtree.h>

Public Member Functions

 DAI_ENUM (UpdateType, HUGIN, SHSH)
 Enumeration of possible JTree updates.
 DAI_ENUM (InfType, SUMPROD, MAXPROD)
 Enumeration of inference variants.
 DAI_ENUM (HeuristicType, MINNEIGHBORS, MINWEIGHT, MINFILL, WEIGHTEDMINFILL)
 Enumeration of elimination cost functions used for constructing the junction tree.

Public Attributes

size_t verbose
 Verbosity (amount of output sent to stderr)
UpdateType updates
 Type of updates.
InfType inference
 Type of inference.
HeuristicType heuristic
 Heuristic to use for constructing the junction tree.
size_t maxmem
 Maximum memory to use in bytes (0 means unlimited)

Detailed Description

Parameters for JTree.

Member Function Documentation

dai::JTree::Properties::DAI_ENUM ( UpdateType  ,
HUGIN  ,
SHSH   
)

Enumeration of possible JTree updates.

There are two types of updates:

  • HUGIN similar to those in HUGIN
  • SHSH Shafer-Shenoy type
dai::JTree::Properties::DAI_ENUM ( InfType  ,
SUMPROD  ,
MAXPROD   
)

Enumeration of inference variants.

There are two inference variants:

  • SUMPROD Sum-Product
  • MAXPROD Max-Product (equivalent to Min-Sum)
dai::JTree::Properties::DAI_ENUM ( HeuristicType  ,
MINNEIGHBORS  ,
MINWEIGHT  ,
MINFILL  ,
WEIGHTEDMINFILL   
)

Enumeration of elimination cost functions used for constructing the junction tree.

The cost of eliminating a variable can be (

See Also
[KoF09], page 314)):
  • MINNEIGHBORS the number of neighbors it has in the current adjacency graph;
  • MINWEIGHT the product of the number of states of all neighbors in the current adjacency graph;
  • MINFILL the number of edges that need to be added to the adjacency graph due to the elimination;
  • WEIGHTEDMINFILL the sum of weights of the edges that need to be added to the adjacency graph due to the elimination, where a weight of an edge is the produt of weights of its constituent vertices. The elimination sequence is chosen greedily in order to minimize the cost.

Member Data Documentation

size_t dai::JTree::Properties::verbose

Verbosity (amount of output sent to stderr)

UpdateType dai::JTree::Properties::updates

Type of updates.

InfType dai::JTree::Properties::inference

Type of inference.

HeuristicType dai::JTree::Properties::heuristic

Heuristic to use for constructing the junction tree.

size_t dai::JTree::Properties::maxmem

Maximum memory to use in bytes (0 means unlimited)


The documentation for this struct was generated from the following file: