ve
Class DecisionNetwork
java.lang.Object
|
+--ve.DecisionNetwork
- public class DecisionNetwork
- extends java.lang.Object
A DecisionNetwork contains a tuple of Variables (each with a domain)
and a set of conditional probability tables, each represented as a
FactorCPT
.
Constructor Summary |
DecisionNetwork(java.lang.String s,
int maxVariables)
constructs a decision network from the extended CIspace
representation. |
Methods inherited from class java.lang.Object |
clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait |
DecisionNetwork
public DecisionNetwork(java.lang.String s,
int maxVariables)
- constructs a decision network from the extended CIspace
representation. The extended representation lets us have
indexes or names as the parents. It assumes that the names are
not numbers (otherwise it assumes they are the indexes).
- Parameters:
s
- string representation of the decision networkmaxVariables
- the maximum number of variables allowed
getVariables
public Variable[] getVariables()
- get the variables of the decision network, in order. The
ordering of indexes to that array should be consistent with the
ordering of the variables, but the indexes don't have to
coincide with the IDs of the variables. The value
getVariables().length does not give the actual number of
variables; use getNumVariables to determine the number of
variables.
getProbFactors
public Factor[] getProbFactors()
- return the conditional probability tables asssociated with the
variables. Decision Variables do not have CPTs, so
there are generally fewer factors than variables. The value
getProbFactors().length does not give the actual number of
factors; use getNumProbFactors to determine the number of
factors.
getNumVariables
public int getNumVariables()
getNumProbFactors
public int getNumProbFactors()
getStringToVar
public java.util.HashMap getStringToVar()
- returns a map from the string to a variable.