Package com.bayesserver.analysis

Class ParameterTuning

java.lang.Object

com.bayesserver.analysis.ParameterTuning

public final class ParameterTuning
extends Object

Calculates how a parameter can be updated so that the resulting value of a hypothesis is within a given range.

Method Summary

Modifier and Type	Method	Description
static ParameterTuningOneWay	oneWayDifference(SensitivityFunctionOneWay f1, SensitivityFunctionOneWay f2, Interval<Double> constraint)	Given a pair of sensitivity functions (evaluated on the same parameter and evidence but different hypotheses), determines how the parameter under consideration can be altered so that the difference between the hypothesis probabilities P(h1|e) - P(h2|e) is within a given range.
static ParameterTuningOneWay	oneWayRatio(SensitivityFunctionOneWay f1, SensitivityFunctionOneWay f2, Interval<Double> constraint)	Given a pair of sensitivity functions (evaluated on the same parameter and evidence but different hypotheses), determines how the parameter under consideration can be altered so that the ratio between the hypothesis probabilities P(h1|e) / P(h2|e) is within a given range.
static ParameterTuningOneWay	oneWaySimple(SensitivityFunctionOneWay f, Interval<Double> constraint)	Given a sensitivity function, determines how the parameter under consideration can be altered so that the resulting value of the hypothesis is within a given range.

Methods inherited from class java.lang.Object

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

Method Detail

oneWaySimple

public static ParameterTuningOneWay oneWaySimple(SensitivityFunctionOneWay f,
                                                 Interval<Double> constraint)
                                          throws ConstraintSatisfiedException,
                                                 ConstraintNotSatisfiedException

Given a sensitivity function, determines how the parameter under consideration can be altered so that the resulting value of the hypothesis is within a given range.

Parameters:

f - The sensitivity function.

constraint - A goal/range we would like the hypothesis probability to lie within.

Returns:

The interval for the parameter which satisfies the desired hypothesis constraint/range.

Throws:

ConstraintSatisfiedException - Raised if the constraint is already satisfied.

ConstraintNotSatisfiedException - Raised if the constraint cannot be satisfied.

oneWayDifference

public static ParameterTuningOneWay oneWayDifference(SensitivityFunctionOneWay f1,
                                                     SensitivityFunctionOneWay f2,
                                                     Interval<Double> constraint)
                                              throws ConstraintSatisfiedException,
                                                     ConstraintNotSatisfiedException

Given a pair of sensitivity functions (evaluated on the same parameter and evidence but different hypotheses), determines how the parameter under consideration can be altered so that the difference between the hypothesis probabilities P(h1|e) - P(h2|e) is within a given range.

Parameters:

f1 - The first sensitivity function.

f2 - The second sensitivity function.

constraint - A goal/range we would like P(h1|e) - P(h2|e) to lie within.

Returns:

The interval for the parameter which satisfies the goal.

Throws:

ConstraintSatisfiedException - Raised if the constraint is already satisfied.

ConstraintNotSatisfiedException - Raised if the constraint cannot be satisfied.

oneWayRatio

public static ParameterTuningOneWay oneWayRatio(SensitivityFunctionOneWay f1,
                                                SensitivityFunctionOneWay f2,
                                                Interval<Double> constraint)
                                         throws ConstraintSatisfiedException,
                                                ConstraintNotSatisfiedException

Given a pair of sensitivity functions (evaluated on the same parameter and evidence but different hypotheses), determines how the parameter under consideration can be altered so that the ratio between the hypothesis probabilities P(h1|e) / P(h2|e) is within a given range.

Parameters:

f1 - The first sensitivity function.

f2 - The second sensitivity function.

constraint - A goal/range we would like P(h1|e) / P(h2|e) to lie within.

Returns:

The interval for the parameter which satisfies the goal.

Throws:

ConstraintSatisfiedException - Raised if the constraint is already satisfied.

ConstraintNotSatisfiedException - Raised if the constraint cannot be satisfied.