To run the Variable Elimination algorithm, create a VE Object and pass in the elimination ordering as arguments. The arguments are of VariableType, an enumeration type and represents the nodes in the Bayes Net.
VE ve = new VE(VariableType.M, VariableType.S, VariableType.C, VariableType.B, VariableType.I);To set an evidence observation before computing a query, simply use the follow command:
ve.setEvidenceObservation(VariableType.S, true);To actually compute a query, simply call the following method:
double result = ve.computeQuery(VariableType.M);That computes the probability, given any evidence supplied, of M being true. To know the probability of M being false, simply take 1 - result.
Finally, if you would like to see a more detailed breakdown of the steps involved and the states of the Factor tables when computing a query then turn on debugging:
ve.turnOnDebugOutput();That will log to the console the state of all the Factor tables for each elimination.
To run the MCMC algorithm, it is very similar to Variable Elimination. Simply create a MCMC Object and pass in the number of iterations per MCMC, N and the number of times MCMC should be run, M. Here MCMC is setup to run for 100 iterations, and will be repeated 1000 times and an average taken.
MCMC mc = new MCMC(100, 1000);To set an evidence observation before computing a query, simply use the following command stating the variable observed, and the actual observed value.
mc.setEvidenceObservation(VariableType.S, true);To actually compute a query, simply call the following method:
double result = mc.computeQuery(VariableType.M);That computes the probability, given any evidence supplied, of M being true. To know the probability of M being false, simply take 1 - result.