broadwick.montecarlo

Interface MonteCarloResults

All Known Implementing Classes:

MonteCarloDefaultResults, Poison

public interface MonteCarloResults

This interface encapsulates the results from a Monte Carlo run. By Default a Monte Carlo object will run simulations and use implementations of this class to store the results. It does do by adding the simulation results to the Samples object that is used to obtain an expected value for the simulation.

Method Summary

Modifier and Type	Method and Description
double	getExpectedValue() Get the score for the Monte Carlo simulation.
Samples	getSamples() Get the results of the Monte Carlo simulations as a Samples object.
MonteCarloResults	join(MonteCarloResults results) Join another Monte Carlo results object to this one.
void	reset() Clear the results of the MonteCarloResults object.
String	toCsv() Get a CSV string of the Monte Carlo results stored in this class.

Method Detail

getExpectedValue

double getExpectedValue()

Get the score for the Monte Carlo simulation. The Acceptor interface and it's implementing classes require a known value in this class for their analysis, e.g. this method may return the likelihood of the simulation and the Acceptor class (e.g. MetropolisHastingsAcceptor) will use the McResults.getExpectedValue() method to determine whether or not to accept this proposed step.

Returns:

the value of the score.

getSamples

Samples getSamples()

Get the results of the Monte Carlo simulations as a Samples object.

Returns:

A samples object that contains the results of each Monte Carlo Simulation.

toCsv

String toCsv()

Get a CSV string of the Monte Carlo results stored in this class. This will be used for logging and for output to file and can contain statistics about the results e.g. the mean, std deviation etc. as appropriate.

Returns:

a CSV representation of the results.

join

MonteCarloResults join(MonteCarloResults results)

Join another Monte Carlo results object to this one. When running a Monte Carlo simulation it is often necessary to append the results from a simulation to a 'master' results object.

Parameters:

results - the results object to join.

Returns:

the new/updated results object.

reset

void reset()

Clear the results of the MonteCarloResults object. This will be called by several algorithms (e.g. MCMC) to clear any cached results that should be cleared between runs.