keel.Algorithms.Rule_Learning.C45RulesSA

Class Cut

java.lang.Object

keel.Algorithms.Rule_Learning.C45RulesSA.Cut

public class Cut
extends java.lang.Object

Class to implement the calculus of the cut point

Since:

JDK1.2

Version:

1.1

Author:

Written by Antonio Alejandro Tortosa (University of Granada) 01/07/2008, Modified by Xavi Solé (La Salle, Ramón Llull University - Barcelona) 16/12/2008

Field Summary

Fields

Modifier and Type	Field and Description
protected Classification	classification Classification of class values.
protected int	numSubsets Number of subsets.

Constructor Summary

Constructors

Constructor and Description
Cut(Classification dist) Function to use when no cut is necessary.
Cut(int index, int nObj, double weights) Function to initialize the cut model.

Method Summary

Modifier and Type	Method and Description
int	attributeIndex() Returns the index of the attribute to cut on.
boolean	checkModel() Function to check if generated model is valid.
Classification	classification() Returns the classification created by the model.
void	classify(MyDataset trainItemsets) Function to create the cut point.
double	classProbability(int classIndex, Itemset itemset, int subset) Function to compute the probability for itemset.
MyDataset[]	cutDataset(MyDataset data) Function to cut the dataset in subsets.
static java.lang.String	doubleToString(double value, int afterDecimalPoint) Function to round a double and converts it into String.
static java.lang.String	doubleToString(double value, int width, int afterDecimalPoint) Function to round a double and converts it into String.
double	gainRatioCutCrit(Classification values, double totalnoInst, double numerator) Function to compute the gain ratio.
double	getCutPoint() Returns the cut point where the node will be divided in its children.
double	getGainRatio() Returns the gain ratio for the cut.
double	getInfoGain() Returns information gain for the generated cut.
double	infoGainCutCrit(Classification values, double totalNoInst, double oldEnt) Function to compute the information gain.
java.lang.String	label(int index, MyDataset data) Function to print label for subset index of itemsets.
java.lang.String	leftSide(MyDataset data) Function to print left side of condition.
protected double	logFunc(double num) Returns the log2
double	newEntropy(Classification values) Function to compute entropy of classification after cutting.
int	numSubsets() Returns the number of created subsets for the cut.
double	oldEntropy(Classification values) Function to compute entropy of classification before cutting.
void	resetClassification(MyDataset data) Function to reset the classification of the model.
java.lang.String	rightSide(int index, MyDataset data) Function to print the condition satisfied by itemsets in a subset.
void	setCutPoint(MyDataset allItemsets) Function to set the cut point.
double[]	weights(Itemset itemset) Returns weights if itemset is assigned to more than one subset, null otherwise.
int	whichSubset(Itemset itemset) Returns index of subset itemset is assigned to.

Methods inherited from class java.lang.Object

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

Field Detail

classification

protected Classification classification

Classification of class values.

numSubsets

protected int numSubsets

Number of subsets.

Constructor Detail

Cut

public Cut(int index,
           int nObj,
           double weights)

Function to initialize the cut model.

Parameters:

index - The attribute index.

nObj - Minimum number of itemsets.

weights - The weight of all the itemsets.

Cut

public Cut(Classification dist)

Function to use when no cut is necessary.

Parameters:

dist - Distribution of values per class.

Method Detail

classify

public void classify(MyDataset trainItemsets)
              throws java.lang.Exception

Function to create the cut point.

Parameters:

trainItemsets - The dataset to classify.

Throws:

java.lang.Exception - If the classification cannot be made.

classProbability

public final double classProbability(int classIndex,
                                     Itemset itemset,
                                     int subset)

Function to compute the probability for itemset.

Parameters:

classIndex - The index of the class.

itemset - The itemset.

subset - The index of the subset.

Returns:

The probability computed.

setCutPoint

public final void setCutPoint(MyDataset allItemsets)

Function to set the cut point.

Parameters:

allItemsets - The dataset used for the cut.

cutDataset

public final MyDataset[] cutDataset(MyDataset data)
                             throws java.lang.Exception

Function to cut the dataset in subsets.

Parameters:

data - The dataset to cut.

Returns:

All the datasets created.

Throws:

java.lang.Exception - If the dataset cannot be cut.

resetClassification

public void resetClassification(MyDataset data)
                         throws java.lang.Exception

Function to reset the classification of the model.

Parameters:

data - The new dataset used.

Throws:

java.lang.Exception - If the classification cannot be reset.

weights

public final double[] weights(Itemset itemset)

Returns weights if itemset is assigned to more than one subset, null otherwise.

Parameters:

itemset - The itemset.

Returns:

weights if itemset is assigned to more than one subset, null otherwise.

whichSubset

public final int whichSubset(Itemset itemset)

Returns index of subset itemset is assigned to.

Parameters:

itemset - The itemset.

Returns:

index of subset itemset is assigned to.

checkModel

public final boolean checkModel()

Function to check if generated model is valid.

Returns:

True if the model is valid. False otherwise.

classification

public final Classification classification()

Returns the classification created by the model.

Returns:

the classification created by the model.

numSubsets

public final int numSubsets()

Returns the number of created subsets for the cut.

Returns:

the number of created subsets for the cut.

gainRatioCutCrit

public final double gainRatioCutCrit(Classification values,
                                     double totalnoInst,
                                     double numerator)

Function to compute the gain ratio.

Parameters:

values - The classification used to compute the gain ratio.

totalnoInst - Number of itemsets.

numerator - The information gain.

Returns:

The gain ratio for the classification.

infoGainCutCrit

public final double infoGainCutCrit(Classification values,
                                    double totalNoInst,
                                    double oldEnt)

Function to compute the information gain.

Parameters:

values - The classification used to compute the information gain.

totalNoInst - Number of itemsets.

oldEnt - The value for the entropy before cutting.

Returns:

The information gain.

oldEntropy

public final double oldEntropy(Classification values)

Function to compute entropy of classification before cutting.

Parameters:

values - The classification used to compute the entropy before cutting.

Returns:

The entropy for the classification before cutting.

newEntropy

public final double newEntropy(Classification values)

Function to compute entropy of classification after cutting.

Parameters:

values - The classification used to compute the entropy after cutting.

Returns:

The entropy for the classification after cutting.

logFunc

protected final double logFunc(double num)

Returns the log2

Parameters:

num - The number to compute the log2.

Returns:

the log2

getInfoGain

public final double getInfoGain()

Returns information gain for the generated cut.

Returns:

information gain for the generated cut.

getGainRatio

public final double getGainRatio()

Returns the gain ratio for the cut.

Returns:

the gain ratio for the cut.

leftSide

public final java.lang.String leftSide(MyDataset data)

Function to print left side of condition.

Parameters:

data - The dataset.

Returns:

The name of the attribute used in the cut.

rightSide

public final java.lang.String rightSide(int index,
                                        MyDataset data)

Function to print the condition satisfied by itemsets in a subset.

Parameters:

index - The index of the value.

data - The dataset.

Returns:

The value for the attribute of the cut.

label

public final java.lang.String label(int index,
                                    MyDataset data)

Function to print label for subset index of itemsets.

Parameters:

index - The index of the subset.

data - The dataset.

Returns:

The label created.

attributeIndex

public final int attributeIndex()

Returns the index of the attribute to cut on.

Returns:

the index of the attribute to cut on.

doubleToString

public static java.lang.String doubleToString(double value,
                                              int afterDecimalPoint)

Function to round a double and converts it into String.

Parameters:

value - The value to print.

afterDecimalPoint - Number of decimals positions.

Returns:

The value with the given number of decimals.

doubleToString

public static java.lang.String doubleToString(double value,
                                              int width,
                                              int afterDecimalPoint)

Function to round a double and converts it into String.

Parameters:

value - The value to print.

width - The width that must have the string generated.

afterDecimalPoint - Number of decimals positions.

Returns:

The value with the given number of decimals.

getCutPoint

public double getCutPoint()

Returns the cut point where the node will be divided in its children.

Returns:

the cut point.