iceCube.uhe.analysis

クラス ModelParaPvalueCalculatorByPoissonBinnedLH

java.lang.Object

iceCube.uhe.analysis.ModelParaPvalueCalculatorByPoissonBinnedLH

public class ModelParaPvalueCalculatorByPoissonBinnedLH
extends java.lang.Object

Calculate the p-value on given neutrino model parameters based on the binned Poisson likelihood for estimating the confidence interval. The PoissonBinnedLikelihoodCalculator :

   PoissonBinnedLikelihoodCalculator  calBG     (for the atmospheric background)
   PoissonBinnedLikelihoodCalculator  calSignal   (for the neutrino model such as GZK with the parameters to be tested)
   PoissonBinnedLikelihoodCalculator  calNuisanceSingal (for the nuisance signal - like E^-2 against the GZK test) 
   

are generated by the correponding binned data from the files and then executes the methods provied by ModelTestByPoissonBinnedLikelihoodFactory. The signal hypothesis with all the other values of parameters are stored in the list container calSignalSetsList. The example for using this class - The constraints on the source evolution parameters (m,Zmax). The likeliihood ratio

 
   alternative (calBG + model stored in calSignalSetsList (+ nuisance ) with parameters to maximize Likelihood)/
   null hypothesis(calBG + calSignal with a given neutrino parameter (+ nuisance to maximize likelihood))
   

is used for the test-statstic

フィールドの概要

Fields

Modifier and Type	Field and 説明
PoissonBinnedLikelihoodCalculator	calBG the poisson bin data container for BG
PoissonBinnedLikelihoodCalculator	calNuisanceSignal container for Nuisance Signal for the profole likelihood
PoissonBinnedLikelihoodCalculator	calSignal the container for Signal with the set of parameter values to be tested
protected java.util.List	calSignalSetsList List to store the PoissonBinnedLikelihoodCalculator calSignal with various model parameters
protected PoissonBinnedLikelihoodCalculator	calSignalToMaxLH container of the signal with the model parameters that maximized likelihood
protected boolean	includeNuisance
static int	runTimesDefault default number of runs in the replica experiment to calculate the pdf of the likelihood ratio
protected ModelTestByPoissonBinnedLikelihoodFactory	testFactory The test factory

コンストラクタの概要

Constructors

Constructor and 説明
ModelParaPvalueCalculatorByPoissonBinnedLH(PoissonBinnedLikelihoodCalculator calBG, PoissonBinnedLikelihoodCalculator calSignal, java.util.List calSignalSetsList) constructor for no nuislance case: PoissonBinnedLikelihoodCalculator calBG :for the atmospheric background PoissonBinnedLikelihoodCalculator calSignal :for the neutrino model with the parmeters to be tested List calSignalSetsList : A container to store PoissonBinnedLikelihoodCalculator objects of the signal hypothesis but with all the parameter values allowd in their range.
ModelParaPvalueCalculatorByPoissonBinnedLH(PoissonBinnedLikelihoodCalculator calBG, PoissonBinnedLikelihoodCalculator calSignal, java.util.Map calSignalMap) constructor for no nuislance case: PoissonBinnedLikelihoodCalculator calBG :for the atmospheric background PoissonBinnedLikelihoodCalculator calSignal :for the neutrino model with the parmeters to be tested Map calSignalMap : A container to store PoissonBinnedLikelihoodCalculator objects of the signal hypothesis but with all the parameter values allowd in their range.
ModelParaPvalueCalculatorByPoissonBinnedLH(PoissonBinnedLikelihoodCalculator calBG, PoissonBinnedLikelihoodCalculator calSignal, PoissonBinnedLikelihoodCalculator calNuisanceSignal, java.util.List calSignalSetsList) constructor: PoissonBinnedLikelihoodCalculator calBG :for the atmospheric background PoissonBinnedLikelihoodCalculator calSignal :for the neutrino model with the parmeters to be tested PoissonBinnedLikelihoodCalculator calNuisanceSingal:for the nuisance signal - like E^-2 against the GZK test List calSignalSetsList : A container to store PoissonBinnedLikelihoodCalculator objects of the signal hypothesis but with all the parameter values allowd in their range.
ModelParaPvalueCalculatorByPoissonBinnedLH(PoissonBinnedLikelihoodCalculator calBG, PoissonBinnedLikelihoodCalculator calSignal, PoissonBinnedLikelihoodCalculator calNuisanceSignal, java.util.Map calSignalMap) constructor: PoissonBinnedLikelihoodCalculator calBG :for the atmospheric background PoissonBinnedLikelihoodCalculator calSignal :for the neutrino model with the parmeters to be tested PoissonBinnedLikelihoodCalculator calNuisanceSingal:for the nuisance signal - like E^-2 against the GZK test Map calSignalMap : A container to store PoissonBinnedLikelihoodCalculator objects of the signal hypothesis but with all the parameter values allowd in their range.

メソッドの概要

Methods

Modifier and Type	Method and 説明
void	buildCalSignalSetsListFromMap(java.util.Map calSignalMap) build the list to store PoissonBinnedLikelihoodCalculator calSignal with various model parameters from Map(Key=PoissonBinnedLikelihoodCalculator).
double	getLLHRatio(boolean runReplicaExperiment) Scan all the PoissonBinnedLikelihoodCalculators stored in List calSignalSetsList to calculate the LLH.
java.util.ListIterator	getllhRatioIterator() return the list iterator to access the list container of the llhRatio data obtained by running replica experiments with the method makeCollectionOfLogLikelihoodRatio()
PoissonBinnedLikelihoodCalculator	getModelToMaximizeLH() Return the PoissonBinnedLikelihoodCalculator to contain the model that maximized the poisson binned likelihood.
double	getPvalue(double llhRatioObserved) Calculate the p-value of the model parameters of calSignal.
double	getPvalue(double llhRatioObserved, int trialTimes) Calculate the p-value of the model parameters of calSignal.
java.util.ListIterator	getSignalSetsListIterator() return the list iterator to access the signal poisson binned data with various model parameters.
static void	main(java.lang.String[] args) A simple main method for debugging
void	makeCollectionOfLogLikelihoodRatio(int runTimes) Calculate collections of the log likelihhood ratio alternative (calBG + model stored in calSignalSetsList (+ nuisance ) with parameters to maximize Likelihood)/ null hypothesis(calBG + calSignal with a given neutrino parameter (+ nuisance to maximize likelihood)) by running replica experiments.
void	useTheFullSearchToMaximizeLLH() Use the full search for the parameters to maximize LLH in the method getLLHRatio() or getPvalue().
void	useTheQuickSearchToMaximizeLLH() Use the quick search for the parameters to maximize LLH in the method getLLHRatio() or getPvalue().

クラス {0} から継承されたメソッド java.lang.Object

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

フィールドの詳細

calBG

public PoissonBinnedLikelihoodCalculator calBG

the poisson bin data container for BG

calSignal

public PoissonBinnedLikelihoodCalculator calSignal

the container for Signal with the set of parameter values to be tested

calNuisanceSignal

public PoissonBinnedLikelihoodCalculator calNuisanceSignal

container for Nuisance Signal for the profole likelihood

calSignalToMaxLH

protected PoissonBinnedLikelihoodCalculator calSignalToMaxLH

container of the signal with the model parameters that maximized likelihood

testFactory

protected ModelTestByPoissonBinnedLikelihoodFactory testFactory

The test factory

runTimesDefault

public static final int runTimesDefault

default number of runs in the replica experiment to calculate the pdf of the likelihood ratio

関連項目:

定数フィールド値

includeNuisance

protected boolean includeNuisance

calSignalSetsList

protected java.util.List calSignalSetsList

List to store the PoissonBinnedLikelihoodCalculator calSignal with various model parameters

コンストラクタの詳細

ModelParaPvalueCalculatorByPoissonBinnedLH

public ModelParaPvalueCalculatorByPoissonBinnedLH(PoissonBinnedLikelihoodCalculator calBG,
                                          PoissonBinnedLikelihoodCalculator calSignal,
                                          PoissonBinnedLikelihoodCalculator calNuisanceSignal,
                                          java.util.List calSignalSetsList)

constructor:

   PoissonBinnedLikelihoodCalculator  calBG            :for the atmospheric background
   PoissonBinnedLikelihoodCalculator  calSignal        :for the neutrino model with the parmeters to be tested
   PoissonBinnedLikelihoodCalculator  calNuisanceSingal:for the nuisance signal - like E^-2 against the GZK test
   List  calSignalSetsList   :  A container to store PoissonBinnedLikelihoodCalculator objects
                               of the signal hypothesis but with all the parameter values allowd in their range.
                               It will build the alternative hypothesis.
       

ModelParaPvalueCalculatorByPoissonBinnedLH

public ModelParaPvalueCalculatorByPoissonBinnedLH(PoissonBinnedLikelihoodCalculator calBG,
                                          PoissonBinnedLikelihoodCalculator calSignal,
                                          java.util.List calSignalSetsList)

constructor for no nuislance case:

   PoissonBinnedLikelihoodCalculator  calBG            :for the atmospheric background
   PoissonBinnedLikelihoodCalculator  calSignal        :for the neutrino model with the parmeters to be tested
   List  calSignalSetsList   :  A container to store PoissonBinnedLikelihoodCalculator objects
                               of the signal hypothesis but with all the parameter values allowd in their range.
                               It will build the alternative hypothesis.
       

ModelParaPvalueCalculatorByPoissonBinnedLH

public ModelParaPvalueCalculatorByPoissonBinnedLH(PoissonBinnedLikelihoodCalculator calBG,
                                          PoissonBinnedLikelihoodCalculator calSignal,
                                          PoissonBinnedLikelihoodCalculator calNuisanceSignal,
                                          java.util.Map calSignalMap)

constructor:

   PoissonBinnedLikelihoodCalculator  calBG            :for the atmospheric background
   PoissonBinnedLikelihoodCalculator  calSignal        :for the neutrino model with the parmeters to be tested
   PoissonBinnedLikelihoodCalculator  calNuisanceSingal:for the nuisance signal - like E^-2 against the GZK test
   Map  calSignalMap   :  A container to store PoissonBinnedLikelihoodCalculator objects
                          of the signal hypothesis but with all the parameter values allowd in their range.
                          It will build the alternative hypothesis. The PoissonBinnedLikelihoodCalculator objects
                          should be stored as the key (not value) member of the map.
       

ModelParaPvalueCalculatorByPoissonBinnedLH

public ModelParaPvalueCalculatorByPoissonBinnedLH(PoissonBinnedLikelihoodCalculator calBG,
                                          PoissonBinnedLikelihoodCalculator calSignal,
                                          java.util.Map calSignalMap)

constructor for no nuislance case:

   PoissonBinnedLikelihoodCalculator  calBG            :for the atmospheric background
   PoissonBinnedLikelihoodCalculator  calSignal        :for the neutrino model with the parmeters to be tested
   Map  calSignalMap   :  A container to store PoissonBinnedLikelihoodCalculator objects
                          of the signal hypothesis but with all the parameter values allowd in their range.
                          It will build the alternative hypothesis. The PoissonBinnedLikelihoodCalculator objects
                          should be stored as the key (not value) member of the map.
       

メソッドの詳細

useTheQuickSearchToMaximizeLLH

public void useTheQuickSearchToMaximizeLLH()

Use the quick search for the parameters to maximize LLH in the method getLLHRatio() or getPvalue(). Scan only the models with the parameters yielding the total expected rate closer to the total observed number of events.

useTheFullSearchToMaximizeLLH

public void useTheFullSearchToMaximizeLLH()

Use the full search for the parameters to maximize LLH in the method getLLHRatio() or getPvalue(). Scan all the models stored in calSignalSetsList. This is the default procedure.

getLLHRatio

public double getLLHRatio(boolean runReplicaExperiment)

Scan all the PoissonBinnedLikelihoodCalculators stored in List calSignalSetsList to calculate the LLH. Then indentify the one that gives a maximal likelihood. If you register the nuisance model in calling the constructor, the profile likelihood is calculated. The likelihood ratio

 
      alternative (calBG + model stored in calSignalSetsList (+ nuisance ) with parameters to maximize Likelihood)/
      null hypothesis(calBG + calSignal with a given neutrino parameter (+ nuisance to maximize likelihood))
      

is returned.

getModelToMaximizeLH

public PoissonBinnedLikelihoodCalculator getModelToMaximizeLH()

Return the PoissonBinnedLikelihoodCalculator to contain the model that maximized the poisson binned likelihood. Run getLLHRatio(boolean runReplicaExperiment) first to scan all the parameters for searching the maximal likelihood.

makeCollectionOfLogLikelihoodRatio

public void makeCollectionOfLogLikelihoodRatio(int runTimes)

Calculate collections of the log likelihhood ratio

 
        alternative (calBG + model stored in calSignalSetsList (+ nuisance ) with parameters to maximize Likelihood)/
        null hypothesis(calBG + calSignal with a given neutrino parameter (+ nuisance to maximize likelihood))
        

by running replica experiments.

getPvalue

public double getPvalue(double llhRatioObserved,
               int trialTimes)

Calculate the p-value of the model parameters of calSignal. The test statstic is the lilelihood ratio

 
        alternative (calBG + model stored in calSignalSetsList (+ nuisance ) with parameters to maximize Likelihood)/
        null hypothesis(calBG + calSignal with a given neutrino parameter (+ nuisance to maximize likelihood))
        

The p.d.f. of the test statstic is simulated by calling the method makeCollectionOfLogLikelihoodRatio(runTimesDefault) internally. int trialTimes is the trail number to run the pseudo-experiments in the p-value calculation. The acuracy of p-value will be O(1/trialTimes).

getPvalue

public double getPvalue(double llhRatioObserved)

Calculate the p-value of the model parameters of calSignal. The test statstic is the lilelihood ratio

 
        alternative (calBG + model stored in calSignalSetsList (+ nuisance ) with parameters to maximize Likelihood)/
        null hypothesis(calBG + calSignal with a given neutrino parameter (+ nuisance to maximize likelihood))
        

The p.d.f. of the test statstic is simulated by calling the method makeCollectionOfLogLikelihoodRatio(runTimesDefault) internally. int runTimes is the trail number to run the pseudo-experiments in the p-value calculation. The default runTime number is used in the p-value calculation.

getSignalSetsListIterator

public java.util.ListIterator getSignalSetsListIterator()

return the list iterator to access the signal poisson binned data with various model parameters.

getllhRatioIterator

public java.util.ListIterator getllhRatioIterator()

return the list iterator to access the list container of the llhRatio data obtained by running replica experiments with the method makeCollectionOfLogLikelihoodRatio()

buildCalSignalSetsListFromMap

public void buildCalSignalSetsListFromMap(java.util.Map calSignalMap)

build the list to store PoissonBinnedLikelihoodCalculator calSignal with various model parameters from Map(Key=PoissonBinnedLikelihoodCalculator). It internally builds calSignalSetsList, the container to store PoissonBinnedLikelihoodCalculator objects of the signal hypothesis but with all the parameter values allowd in their range. This is an utility method.

main

public static void main(java.lang.String[] args)
                 throws java.io.IOException

A simple main method for debugging

例外:

java.io.IOException