Package edu.cmu.tetrad.study

Class RBExperiments

java.lang.Object

edu.cmu.tetrad.study.RBExperiments

public class RBExperiments
extends Object

RBExperiments class.

Version:

$Id: $Id

Author:

josephramsey

Constructor Summary

Constructors

Constructor	Description
RBExperiments()	Constructor for RBExperiments.

Method Summary

Modifier and Type	Method	Description
void	experiment(String modelName, int numCases, int numModels, int numBootstrapSamples, double alpha, double numLatentConfounders, boolean threshold1, boolean threshold2, double lower, double upper, String filePath, int round)	Performs an experiment to estimate the structure and parameters of a Bayesian network using various methods.
static void	main(String[] args)	Main method for executing the RBExperiments class and running a series of experiments.

Methods inherited from class java.lang.Object

equals, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Constructor Details

RBExperiments

public RBExperiments()

Constructor for RBExperiments.

Method Details

main

public static void main(String[] args)
                 throws IOException

Main method for executing the RBExperiments class and running a series of experiments.

Parameters:

args - the command line arguments. These arguments can be used to set various parameters for the experiments. Possible options include: -c (ignored) : a flag to indicate that the input arguments are given in a compact format. -i (ignored) : a flag to indicate that the input arguments are given as individual arguments. -lv [value] : the value of the number of latent confounders. -bs [value] : the number of bootstrap samples. -alpha [value] : the significance level (alpha) for statistical tests. -m [value] : the number of models to consider. -net [value] : the name of the model. -t1 [value] : a flag to indicate whether threshold 1 should be used. -t2 [value] : a flag to indicate whether threshold 2 should be used. -low [value] : the lower threshold value for creating a dependency filtering dataset. -up [value] : the upper threshold value for creating a dependency filtering dataset. -out [value] : the directory to save the results to. -data [value] : the path to the data directory.

Throws:

IOException - if there is an error reading or writing data.

experiment

public void experiment(String modelName,
 int numCases,
 int numModels,
 int numBootstrapSamples,
 double alpha,
 double numLatentConfounders,
 boolean threshold1,
 boolean threshold2,
 double lower,
 double upper,
 String filePath,
 int round)

Performs an experiment to estimate the structure and parameters of a Bayesian network using various methods.

Parameters:

modelName - the name of the model for the Bayesian network

numCases - the number of cases to simulate

numModels - the number of models to run the experiment for

numBootstrapSamples - the number of bootstrap samples to use for creating empirical data

alpha - the significance level for the chi-squared test

numLatentConfounders - the percentage of variables to be set as latent confounders

threshold1 - a flag indicating whether to use threshold 1 during the RB search

threshold2 - a flag indicating whether to use threshold 2 during the RB search

lower - the lower threshold value for the RB search

upper - the upper threshold value for the RB search

filePath - the path to the directory for storing the experiment results

round - the round of the experiment