org.jbpm.sim.report
Class StatUtil

java.lang.Object
  extended by org.jbpm.sim.report.StatUtil

public class StatUtil
extends java.lang.Object

Statistic utils for calculating confidence intervals or sample sizes to satisfy given confidence levels. More information can be found in the statistics chapter of my master thesis.

Author:
bernd.ruecker@camunda.com

Constructor Summary
StatUtil()
           
 
Method Summary
static double getConfidenceIntervalSummand(double confidenceLevel, double standardDerivation, double sampleSize)
          calculating the confidence interval summand, which must be added/subtracted from the sample mean to get the confidence interval (mean +/- confidenceIntervalSummand).
static double getInvCdfParam(double confidenceLevel)
           
static double getRequiredSampleSize(double confidenceLevel, double halfConfidenceIntervalWidth, double standardDerivation)
          computes the required sample size to get the given confidence level and the given confidence interval width (NOTE: The parameter should be the half width, because that is the variable normally added/subtracted from the mean).
 
Methods inherited from class java.lang.Object
clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait
 

Constructor Detail

StatUtil

public StatUtil()
Method Detail

getConfidenceIntervalSummand

public static double getConfidenceIntervalSummand(double confidenceLevel,
                                                  double standardDerivation,
                                                  double sampleSize)
calculating the confidence interval summand, which must be added/subtracted from the sample mean to get the confidence interval (mean +/- confidenceIntervalSummand). The confidence interval is calculated with the "Inverse Cumulative Standard Normal Distribution Function" and not the t distribution which requires a sample size of at least 30 to get qualitative results.

Parameters:
confidenceLevel -
standardDerivation -
sampleSize -
Returns:

getRequiredSampleSize

public static double getRequiredSampleSize(double confidenceLevel,
                                           double halfConfidenceIntervalWidth,
                                           double standardDerivation)
computes the required sample size to get the given confidence level and the given confidence interval width (NOTE: The parameter should be the half width, because that is the variable normally added/subtracted from the mean). The standard derivation of a sample must be provided. The confidence interval is calculated with the "Inverse Cumulative Standard Normal Distribution Function" and not the t distribution which requires a sample size of at least 30 to get qualitative results.

Parameters:
confidenceLevel -
halfConfidenceIntervalWidth -
standardDerivation -
Returns:

getInvCdfParam

public static double getInvCdfParam(double confidenceLevel)


Copyright © 2009 JBoss, a division of Red Hat. All Rights Reserved.