com.itemanalysis.psychometrics.irt.equating

Class StockingLordMethod

java.lang.Object

com.itemanalysis.psychometrics.analysis.AbstractMultivariateFunction

com.itemanalysis.psychometrics.irt.equating.StockingLordMethod

All Implemented Interfaces:

LinearTransformation, Uncmin_methods, org.apache.commons.math3.analysis.MultivariateFunction, org.apache.commons.math3.analysis.UnivariateFunction

public class StockingLordMethod
extends AbstractMultivariateFunction
implements LinearTransformation, Uncmin_methods, org.apache.commons.math3.analysis.UnivariateFunction

Constructor Summary

Constructors

Constructor and Description
StockingLordMethod(java.util.LinkedHashMap<java.lang.String,ItemResponseModel> itemFormX, java.util.LinkedHashMap<java.lang.String,ItemResponseModel> itemFormY, DistributionApproximation xDistribution, DistributionApproximation yDistribution, EquatingCriterionType criterion)

Method Summary

Modifier and Type	Method and Description
double	f_to_minimize(double[] x)
double	getF1(double[] coefficient) Criterion function F1 as described in Kim and Kolen (2007).
double	getF2(double[] coefficient) Criterion function F2 as described in Kim and Kolen (2007).
double	getFormXTccAtTheta(double theta)
double	getFormYTccAtTheta(double theta)
double	getIntercept()
org.apache.commons.math3.optim.nonlinear.scalar.ObjectiveFunction	getObjectiveFunction()
org.apache.commons.math3.optim.nonlinear.scalar.ObjectiveFunctionGradient	getObjectiveFunctionGradient()
double	getScale()
double	getTSharpAtTheta(double[] coefficient, double theta)
double	getTStarAtTheta(double[] coefficient, double theta)
void	gradient(double[] x, double[] g)
void	hessian(double[] x, double[][] h)
void	setIntercept(double intercept)
void	setPrecision(int precision)
void	setScale(double scale)
void	setStandardized(boolean standardized) Flag to standardize criterion function.
double	transform(double x)
double	value(double x) Method needed for Brent optimizer when minimizing teh objective function for the Rasch family of models.
double	value(double[] coefficient) Function to be minimized by optimization class Uncmin index starts at 1 argument[1]=B (intercept) equating constant argument[2]=A (slope) equating constant

Methods inherited from class com.itemanalysis.psychometrics.analysis.AbstractMultivariateFunction

gradient, gradientAt, hessianAt, machineEpsilon, partialDerivative

Methods inherited from class java.lang.Object

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

Constructor Detail

StockingLordMethod

public StockingLordMethod(java.util.LinkedHashMap<java.lang.String,ItemResponseModel> itemFormX,
                          java.util.LinkedHashMap<java.lang.String,ItemResponseModel> itemFormY,
                          DistributionApproximation xDistribution,
                          DistributionApproximation yDistribution,
                          EquatingCriterionType criterion)
                   throws org.apache.commons.math3.exception.DimensionMismatchException

Throws:

org.apache.commons.math3.exception.DimensionMismatchException

Method Detail

setStandardized

public void setStandardized(boolean standardized)

Flag to standardize criterion function. This will affect teh value of the crierion function but no the slope and intercept values.

Parameters:

standardized - if true the criterion function is standardized. If not, it is not standardized.

value

public double value(double[] coefficient)

Function to be minimized by optimization class Uncmin index starts at 1 argument[1]=B (intercept) equating constant argument[2]=A (slope) equating constant

Specified by:

value in interface org.apache.commons.math3.analysis.MultivariateFunction

value

public double value(double x)

Method needed for Brent optimizer when minimizing teh objective function for the Rasch family of models. This method is required by the UnivariateFunction interface.

Specified by:

value in interface org.apache.commons.math3.analysis.UnivariateFunction

Parameters:

x - intercept

Returns:

value of teh criterion function.

getObjectiveFunction

public org.apache.commons.math3.optim.nonlinear.scalar.ObjectiveFunction getObjectiveFunction()

getObjectiveFunctionGradient

public org.apache.commons.math3.optim.nonlinear.scalar.ObjectiveFunctionGradient getObjectiveFunctionGradient()

getF1

public double getF1(double[] coefficient)

Criterion function F1 as described in Kim and Kolen (2007).

Parameters:

coefficient - arguments in optimization

Returns:

criterion function rho

getF2

public double getF2(double[] coefficient)

Criterion function F2 as described in Kim and Kolen (2007).

Parameters:

coefficient - arguments in optimization

Returns:

criterion function values

getFormYTccAtTheta

public double getFormYTccAtTheta(double theta)

getFormXTccAtTheta

public double getFormXTccAtTheta(double theta)

getTStarAtTheta

public double getTStarAtTheta(double[] coefficient,
                              double theta)

getTSharpAtTheta

public double getTSharpAtTheta(double[] coefficient,
                               double theta)

setIntercept

public void setIntercept(double intercept)

setScale

public void setScale(double scale)

getIntercept

public double getIntercept()

Specified by:

getIntercept in interface LinearTransformation

getScale

public double getScale()

Specified by:

getScale in interface LinearTransformation

setPrecision

public void setPrecision(int precision)

transform

public double transform(double x)

Specified by:

transform in interface LinearTransformation

f_to_minimize

public double f_to_minimize(double[] x)

Specified by:

f_to_minimize in interface Uncmin_methods

hessian

public void hessian(double[] x,
                    double[][] h)

Specified by:

hessian in interface Uncmin_methods

gradient

public void gradient(double[] x,
                     double[] g)

Specified by:

gradient in interface Uncmin_methods