Core.CBR

Class CBR

java.lang.Object

Core.CBR.CBR

public class CBR
extends java.lang.Object

Field Summary

Fields

Modifier and Type	Field and Description
protected LearningAttributeWeights	learn Decides the weights of the attributes for the current execution.
protected Retain	retain Computes all the steps of the Retain phase of a CBR.
protected Review	review Evaluates the case base in search of any redundant, innecessary or non beneficial cases, and drops them if found.
protected Revise	revise Computes all the steps of the Revise phase of a CBR.
protected Population	test Test population
protected Population	train Train population

Constructor Summary

Constructors

Constructor and Description
CBR(CBR cbr) Copy constructor.
CBR(Population train, Population test, GlobalDistance distancia_global, Selection seleccio, Reuse reuse, Revise revise, Retain retain, java.lang.String classe) Constructor.

Method Summary

Methods

Modifier and Type	Method and Description
void	afegeixFiltrat(ContextReasoning f) Adds a filtering method to this CBR
void	afegeixLearn(LearningAttributeWeights l) Adds a weights learning method to this CBR
void	afegeixRetain(Retain r) Adds a retain method to this CBR.
void	afegeixReuse(Reuse r) Adds a reuse method to this CBR.
void	afegeixReview(Review r) Adds a review method to this CBR.
void	afegeixRevise(Revise r) Adds a revise method to this CBR.
CBR	clone() Creates a clone of this CBR.
void	freeMemory() Clears the CBR data.
java.lang.String	getClasse() Method to get the attribute for wich the CBR attempts classification.
Classificacio	getClassification() Returns the classification of the train cases
ClassificacioParametres	getClassifications() Returns the parameter classifications of the train cases
MatriuDistancies	getDistanceMatrix() Returns the computeDistance matrix.
java.math.BigDecimal[]	getGamma() This method is a remnant of previous attempts to build a MultiAgent CBR Following its trace it's clear that this method always returns null
double	getGmin(java.lang.String[] timestamps, java.lang.Double[] CGMsamples) Function that calculates the minimum glucose value of a given glucose curve of CGM samples (in mg/dl).
java.math.BigDecimal[]	getPesos() Method to get the current attribute weights.
Retain	getRetain() Returns the retain implementation of this CBR
Reuse	getReuse() Returns the reuse implementation of this CBR
Review	getReview() Returns the review implementation of this CBR
Revise	getRevise() Returns the revise implementation of this CBR
SeleccioIndividus	getSeleccioIndividus() Returns the selected individuals in the retrieve phase
Population	getTest() Method to get the test population of this CBR
Population	getTrain() Method to get the train population of this CBR
float[]	getWeights() Returns the current attribute weights of the train population
Attribute[]	learnWeights() Attribute weight learning phase.
Attribute[]	learnWeights(CBR cbr, Attribute[] a) Attribute weight learning phase.
Attribute[]	learnWeightsCyclic()
void	pepperDataNormalization(Population p)
Population	retain() Retain phase of the CBR.
Population	retainPEPPER() Especific retain for PEPPER project
MatriuDistancies	retrieve() First of the 4 main steps of a CBR.
void	retrievePepper() Retrieve method for the PEPPER project.
void	reuse() Reuse phase of the CBR.
void	reusePepper(double bodyWeight, double Gsp) Reuse method for the PEPPER project.
Population	review() Review phase of the CBR.
void	revise() Revise phase of the CBR.
void	setTest(Individual test, Attribute[] attributes) Sets the test population of this CBR to test and its attributes to attributes
void	setTest(Population test) Sets the test population of this CBR to test
void	setTrain(Population train) Sets the train population of this CBR to test
void	setWeights(float[] weights) Modifies the weights of the train and test population of the CBR

Methods inherited from class java.lang.Object

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

Field Detail

train

protected Population train

Train population

test

protected Population test

Test population

revise

protected Revise revise

Computes all the steps of the Revise phase of a CBR. Decides if the current answer to the class is correct, and computes a correction if needed.

retain

protected Retain retain

Computes all the steps of the Retain phase of a CBR. Decides wether a new case should be added to the current case base.

review

protected Review review

Evaluates the case base in search of any redundant, innecessary or non beneficial cases, and drops them if found.

learn

protected LearningAttributeWeights learn

Decides the weights of the attributes for the current execution. Though machine learning processing attempts to set the attribute weights to the best possible values for a better selection of useful cases.

Constructor Detail

CBR

public CBR(Population train,
   Population test,
   GlobalDistance distancia_global,
   Selection seleccio,
   Reuse reuse,
   Revise revise,
   Retain retain,
   java.lang.String classe)

Constructor. Builda a CBR with the specified methods.

Parameters:

train - Population case base.

test - Population cases to decide its class.

distancia_global - Method to computeDistance the global computeDistance between populations

seleccio - Method to select nearest cases from the test population to the train one

reuse - Method to decide the test individuals' class

revise - Method to revise and correct the classification made in reuse

retain - Method to add test useful cases to the train case base

classe - String defining the attribute for wich a classification is attempted

See Also:

GlobalDistance, Selection, Reuse, Revise, Retain

CBR

public CBR(CBR cbr)

Copy constructor. Builds a CBR equal to the parameter.

Parameters:

cbr - CBR this will be a copy of.

Method Detail

getTest

public Population getTest()

Method to get the test population of this CBR

Returns:

Population test

getTrain

public Population getTrain()

Method to get the train population of this CBR

Returns:

Population train

getPesos

public java.math.BigDecimal[] getPesos()

Method to get the current attribute weights.

Returns:

BigDecimal[] attribute weights.

getGamma

public java.math.BigDecimal[] getGamma()

This method is a remnant of previous attempts to build a MultiAgent CBR Following its trace it's clear that this method always returns null

Returns:

gamma (null value)

getClasse

public java.lang.String getClasse()

Method to get the attribute for wich the CBR attempts classification.

Returns:

String with the name of the class attribute.

getDistanceMatrix

public MatriuDistancies getDistanceMatrix()

Returns the computeDistance matrix. Will be null if it hasn't yet been computed.

Returns:

MatriuDistancies computeDistance matrix.

getSeleccioIndividus

public SeleccioIndividus getSeleccioIndividus()

Returns the selected individuals in the retrieve phase

Returns:

SeleccioIndividus selection of the individuals

afegeixFiltrat

public void afegeixFiltrat(ContextReasoning f)

Adds a filtering method to this CBR

Parameters:

f - ContextReasoning filtering method

afegeixReuse

public void afegeixReuse(Reuse r)

Adds a reuse method to this CBR.

Parameters:

r - Reuse implementation of a Reuse method

See Also:

Reuse

afegeixRevise

public void afegeixRevise(Revise r)

Adds a revise method to this CBR.

Parameters:

r - Revise implementation of a Revise method

See Also:

Revise

afegeixRetain

public void afegeixRetain(Retain r)

Adds a retain method to this CBR.

Parameters:

r - Retain implementation of a Retain method.

See Also:

Retain

afegeixReview

public void afegeixReview(Review r)

Adds a review method to this CBR. #param r Review implementation of a Review method.

Parameters:

r -

See Also:

Review

afegeixLearn

public void afegeixLearn(LearningAttributeWeights l)

Adds a weights learning method to this CBR

Parameters:

l -

See Also:

LearningAttributeWeights

learnWeights

public Attribute[] learnWeights(CBR cbr,
                       Attribute[] a)

Attribute weight learning phase. Prepares the necessary data and executes the implemented weight learning method of this CBR over the data of another cbr.

Parameters:

cbr - CBR other cbr.

a - Attribute[] array with other cbr attributes

Returns:

Attribute[] arrray with this cbr attributes with updated weights

learnWeights

public Attribute[] learnWeights()

Attribute weight learning phase. Prepares the necessary data and executes the implemented weight learning method of this CBR.

Returns:

Attribute[] arrray with this cbr attributes with updated weights

learnWeightsCyclic

public Attribute[] learnWeightsCyclic()

retrieve

public MatriuDistancies retrieve()

First of the 4 main steps of a CBR. Here the matrix of global distances between the test and train population is calculated. The result will be a matrix of Float objects. Float[i][j] represents the computeDistance between hte ith test individual and the jth train individual. The matrix will be stored in the directory especified by parameter.

Returns:

MatriuDistancies.

retrievePepper

public void retrievePepper()

Retrieve method for the PEPPER project. It discretizes some specific variables and then calls the traditional Retrieve

getGmin

public double getGmin(java.lang.String[] timestamps,
             java.lang.Double[] CGMsamples)

Function that calculates the minimum glucose value of a given glucose curve of CGM samples (in mg/dl).

Parameters:

timestamps - timestamps of the CGM samples in the following format "yyyy-MM-dd HH:mm:ss"

CGMsamples - CGM samples of blood glucose in (mg/dl)

Returns:

the minimum glucose value if the CGM curve has enough data, otherwise return -1.

pepperDataNormalization

public void pepperDataNormalization(Population p)

reuse

public void reuse()

Reuse phase of the CBR. Executes the reuse implementation of this CBR over the current data.

reusePepper

public void reusePepper(double bodyWeight,
               double Gsp)

Reuse method for the PEPPER project. It adds attribute information to the case according to the computed classification.

Parameters:

bodyWeight -

Gsp -

revise

public void revise()

Revise phase of the CBR. Executes the revise implementation of this CBR over the current data.

retain

public Population retain()

Retain phase of the CBR. Executes the retain implementation of this CBR over the current populations.

Returns:

Population train plus the newly retained cases.

retainPEPPER

public Population retainPEPPER()

Especific retain for PEPPER project

Returns:

review

public Population review()

Review phase of the CBR. Executes the review implementation of this CBR over the current populations.

Returns:

Population train minus discarded cases.

freeMemory

public void freeMemory()

Clears the CBR data.

getClassification

public Classificacio getClassification()

Returns the classification of the train cases

Returns:

Classificacio classification of the train cases

getClassifications

public ClassificacioParametres getClassifications()

Returns the parameter classifications of the train cases

Returns:

ClassificacioParametres parameter classifications of the train cases

getReuse

public Reuse getReuse()

Returns the reuse implementation of this CBR

Returns:

Reuse reuse implementation

getRevise

public Revise getRevise()

Returns the revise implementation of this CBR

Returns:

Revise revise implementation

getRetain

public Retain getRetain()

Returns the retain implementation of this CBR

Returns:

Retain retain implementation

getReview

public Review getReview()

Returns the review implementation of this CBR

Returns:

Review review implementation

setWeights

public void setWeights(float[] weights)

Modifies the weights of the train and test population of the CBR

Parameters:

weights -

getWeights

public float[] getWeights()

Returns the current attribute weights of the train population

Returns:

float[] weight array of the attributes

setTest

public void setTest(Population test)

Sets the test population of this CBR to test

Parameters:

test - Population to be set as this CBR's test population

setTest

public void setTest(Individual test,
           Attribute[] attributes)

Sets the test population of this CBR to test and its attributes to attributes

Parameters:

test - Population to be set as this CBR's test population

attributes - Attribute[] attributes of the test population

setTrain

public void setTrain(Population train)

Sets the train population of this CBR to test

Parameters:

train - Population to be set as this CBR's train population

clone

public CBR clone()

Creates a clone of this CBR.

Overrides:

clone in class java.lang.Object

Returns:

CBR clone of this CBR.