hu.birot.OTKit.learning
Class ConstraintMotionExamples

java.lang.Object
  hu.birot.OTKit.learning.ConstraintMotionExamples

public class ConstraintMotionExamples
extends java.lang.Object

Includes a number of static methods that return instances of ConstraintMotion.

Constructor Summary

ConstraintMotionExamples()

Method Summary

static ConstraintMotion	allDemotion(double plasticity) Demote all loser-preferring constraints by plasticity.
static ConstraintMotion	antiMagri(double plasticity) Reverse of Giorgio Magri's method of updating a hierarchy, based on ranking variable "rank".
static ConstraintMotion	Boersma(double plasticity) Paul Boersma's method of updating a hierarchy, based on ranking variable "rank".
static ConstraintMotion	demotionOnly(double plasticity) Demote highest ranked loser-preferring constraint by w * plasticity, where w is the number of winner-preferring constraints.
static ConstraintMotion	Magri(double plasticity) Giorgio Magri's method of updating a hierarchy, based on ranking variable "rank".
static ConstraintMotion	singleDemotionOnly(double plasticity) Demote highest ranked loser-preferring constraint by plasticity.
static ConstraintMotion	TesarSmolensky() Tesar & Smolensky's method of updating a hierarchy, based on ranking variable "rank".

Methods inherited from class java.lang.Object

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

Constructor Detail

ConstraintMotionExamples

public ConstraintMotionExamples()

Method Detail

TesarSmolensky

public static ConstraintMotion TesarSmolensky()

Tesar & Smolensky's method of updating a hierarchy, based on ranking variable "rank".

Let W be the highest ranked constraint that prefers the winer. Then all loser-preferring constraints ranked higher than W will be assigned a rank equal to W's rank minus 1.

If both the winner and the loser are generated by some hierarchies, and they have different violation marks, then there must be both at least one winner-preferring constraint, as well as at least one loser-preferring constraints ranked higher than W. In other cases, there can be none. If there is no winner-preferring constraint, or if there is no loser-preferring constraint ranked higher than W, then the hierarchy will not be changed (and the return value of method update will be false.

Returns:

a ConstraintMotion implementing Boersma's approach.

Boersma

public static ConstraintMotion Boersma(double plasticity)

Paul Boersma's method of updating a hierarchy, based on ranking variable "rank".

Demote all loser-preferring constraints by plasticity, and promote all winner-preferring constraint by plasticity.

Parameters:

plasticity - Value by which constraints are promoted or demoted.

Returns:

a ConstraintMotion implementing Boersma's approach.

Magri

public static ConstraintMotion Magri(double plasticity)

Giorgio Magri's method of updating a hierarchy, based on ranking variable "rank".

Promote all winner-preferring constraint by plasticity, and demote highest ranked loser-preferring constraint by w * plasticity, where w is the number of winner-preferring constraints.

Parameters:

plasticity - Value by which constraints are promoted or demoted.

Returns:

a ConstraintMotion implementing Magri's approach.

demotionOnly

public static ConstraintMotion demotionOnly(double plasticity)

Demote highest ranked loser-preferring constraint by w * plasticity, where w is the number of winner-preferring constraints.

Parameters:

plasticity - Value by which constraints are demoted.

Returns:

a ConstraintMotion implementing Magri's approach.

singleDemotionOnly

public static ConstraintMotion singleDemotionOnly(double plasticity)

Demote highest ranked loser-preferring constraint by plasticity.

Parameters:

plasticity - Value by which constraints are demoted.

Returns:

a ConstraintMotion implementing Magri's approach.

allDemotion

public static ConstraintMotion allDemotion(double plasticity)

Demote all loser-preferring constraints by plasticity.

Parameters:

plasticity - Value by which constraints are demoted.

Returns:

a ConstraintMotion implementing Magri's approach.

antiMagri

public static ConstraintMotion antiMagri(double plasticity)

Reverse of Giorgio Magri's method of updating a hierarchy, based on ranking variable "rank".

Demote all loser-preferring constraint by plasticity, and promote lowest ranked winer-preferring constraints by l * plasticity, where l is the number of loser-preferring constraints.

Parameters:

plasticity - Value by which constraints are promoted or demoted.

Returns:

a ConstraintMotion implementing the reverse of Magri's approach.