keel.Algorithms.Decision_Trees.M5

Class M5

java.lang.Object

keel.Algorithms.Decision_Trees.M5.M5

public final class M5
extends java.lang.Object

M5 algorithm main class.

Since:

JDK 1.5

Version:

0.1

Author:

Written by Cristobal Romero (Universidad de Córdoba) 10/10/2007

Field Summary

Fields

Modifier and Type	Field and Description
static int	MODEL_LINEAR_REGRESSION Number to represent type of model used (Lineal regression).
static int	MODEL_MODEL_TREE Number to represent type of model used (tree model).
static int	MODEL_REGRESSION_TREE Number to represent type of model used (tree regression).
static Association[]	TAGS_MODEL_TYPES Tags for the model types.

Constructor Summary

Constructors

Constructor and Description
M5()

Method Summary

Modifier and Type	Method and Description
static double	absDev(int attr, M5Instances inst) Returns the absolute deviation value of the instances values of an attribute
void	buildClassifier(M5Instances inst) Construct a model tree by training instances
double	classifyInstance(M5Instance ins) Classifies the given test instance.
static double	correlation(double[] y1, double[] y2, int n) Returns the correlation coefficient of two double vectors
static java.lang.StringBuffer	deleteTrailingZerosAndDot(java.lang.StringBuffer stringBuffer) Deletes the trailing zeros and decimal point in a stringBuffer
static java.lang.String	doubleToStringF(double value, int width, int afterDecimalPoint) Rounds a double and converts it into a formatted right-justified String.
static java.lang.String	doubleToStringG(double value, int width, int precision) Rounds a double and converts it into a formatted right-justified String.
java.util.Enumeration	enumerateMeasures() Returns an enumeration of the additional measure names
static boolean	eqDouble(double a, double b) Tests if two double values are equal to each other
static void	errorMsg(java.lang.String err) Prints error message and exits
static long	floorDouble(double value) Returns the largest (closest to positive infinity) long integer value that is not greater than the argument.
static M5	forName(java.lang.String classifierName, java.lang.String[] options) Allows to construct the M5 classifier calling it by its name and options.
static java.lang.String	getHeader(java.lang.String fileName) Extracts the header of the file given.
static java.lang.String	getHeaderNoData(java.lang.String fileName) Extracts the header of the file given, when the data is not specified.
double	getMeasure(java.lang.String additionalMeasureName) Returns the value of the named measure
SelectedAssociation	getModelType() Get the value of Model.
java.lang.String[]	getOptions() Gets the current settings of the Classifier.
double	getPruningFactor() Get the value of PruningFactor.
boolean	getUseUnsmoothed() Get the value of UseUnsmoothed.
int	getVerbosity() Get the value of Verbosity.
static boolean	hasEnumAttr(M5Instances inst) Tests if enumerated attribute(s) exists in the instances
static boolean	hasMissing(M5Instances inst) Tests if missing value(s) exists in the instances
static java.lang.String	headToString() Prints the head lines of the output
java.util.Enumeration	listOptions()
static void	main(java.lang.String[] argv) Main method for M5' algorithm
double	measureNumLeaves() return the number of leaves in the tree
double	measureNumLinearModels() return the number of linear models
double	measureNumRules() return the number of rules
static long	roundDouble(double value) Rounds a double
static java.lang.String	separatorToString() Prints sepearating line
void	setModelType(SelectedAssociation newMethod) Set the value of Model.
protected static void	setOptions(java.io.StreamTokenizer options) Name: setOptions Sets the option of the execution of the algorithm, reading them from an input file.
void	setOptions(java.lang.String[] options) Parses a given list of options.
void	setPruningFactor(double v) Set the value of PruningFactor.
void	setUseUnsmoothed(boolean v) Set the value of UseUnsmoothed.
void	setVerbosity(int v) Set the value of Verbosity.
static double	smoothenValue(double p, double q, int n, int k) Returns the smoothed values according to the smoothing formula (np+kq)/(n+k)
static double	sqrSum(int attr, M5Instances inst) Returns the squared sum of the instances values of an attribute
static double	stdDev(int attr, M5Instances inst) Returns the standard deviation value of the instances values of an attribute
static double	sum(int attr, M5Instances inst) Returns the sum of the instances values of an attribute
java.lang.String	toString() Converts the output of the training process into a string
static double	variance(int attr, M5Instances inst) Returns the variance value of the instances values of an attribute

Methods inherited from class java.lang.Object

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

Field Detail

MODEL_LINEAR_REGRESSION

public static final int MODEL_LINEAR_REGRESSION

Number to represent type of model used (Lineal regression).

See Also:

Constant Field Values

MODEL_REGRESSION_TREE

public static final int MODEL_REGRESSION_TREE

Number to represent type of model used (tree regression).

See Also:

Constant Field Values

MODEL_MODEL_TREE

public static final int MODEL_MODEL_TREE

Number to represent type of model used (tree model).

See Also:

Constant Field Values

TAGS_MODEL_TYPES

public static final Association[] TAGS_MODEL_TYPES

Tags for the model types.

Constructor Detail

M5

public M5()

Method Detail

setOptions

protected static void setOptions(java.io.StreamTokenizer options)
                          throws java.lang.Exception

Name: setOptions Sets the option of the execution of the algorithm, reading them from an input file.

Parameters:

options - the tokenizer used to tokenize the input file.

Throws:

java.lang.Exception - if the configuration files can not be read correctly.

getHeader

public static java.lang.String getHeader(java.lang.String fileName)

Extracts the header of the file given.

Parameters:

fileName - file's name give.

Returns:

The sring read.

getHeaderNoData

public static java.lang.String getHeaderNoData(java.lang.String fileName)

Extracts the header of the file given, when the data is not specified.

Parameters:

fileName - file's name give.

Returns:

The sring read.

buildClassifier

public final void buildClassifier(M5Instances inst)
                           throws java.lang.Exception

Construct a model tree by training instances

Parameters:

inst - training instances

Throws:

java.lang.Exception - if the classifier can't be built

classifyInstance

public double classifyInstance(M5Instance ins)
                        throws java.lang.Exception

Classifies the given test instance.

Parameters:

ins - the instance to be classified

Returns:

the predicted class for the instance

Throws:

java.lang.Exception - if the instance can't be classified

listOptions

public java.util.Enumeration listOptions()

Returns:

an enumeration of all the available options

setOptions

public void setOptions(java.lang.String[] options)
                throws java.lang.Exception

Parses a given list of options.

Parameters:

options - the list of options as an array of strings

Throws:

java.lang.Exception - if an option is not supported

getOptions

public java.lang.String[] getOptions()

Gets the current settings of the Classifier.

Returns:

an array of strings suitable for passing to setOptions

toString

public final java.lang.String toString()

Converts the output of the training process into a string

Overrides:

toString in class java.lang.Object

Returns:

the converted string

measureNumLinearModels

public double measureNumLinearModels()

return the number of linear models

Returns:

the number of linear models

measureNumLeaves

public double measureNumLeaves()

return the number of leaves in the tree

Returns:

the number leaves in the tree (same as # linear models & # rules)

measureNumRules

public double measureNumRules()

return the number of rules

Returns:

the number of rules (same as # linear models & # leaves in the tree)

enumerateMeasures

public java.util.Enumeration enumerateMeasures()

Returns an enumeration of the additional measure names

Returns:

an enumeration of the measure names

getMeasure

public double getMeasure(java.lang.String additionalMeasureName)

Returns the value of the named measure

Parameters:

additionalMeasureName - the name of the measure to query for its value

Returns:

the value of the named measure

Throws:

java.lang.IllegalArgumentException - if the named measure is not supported

getUseUnsmoothed

public boolean getUseUnsmoothed()

Get the value of UseUnsmoothed.

Returns:

Value of UseUnsmoothed.

setUseUnsmoothed

public void setUseUnsmoothed(boolean v)

Set the value of UseUnsmoothed.

Parameters:

v - Value to assign to UseUnsmoothed.

getPruningFactor

public double getPruningFactor()

Get the value of PruningFactor.

Returns:

Value of PruningFactor.

setPruningFactor

public void setPruningFactor(double v)

Set the value of PruningFactor.

Parameters:

v - Value to assign to PruningFactor.

getModelType

public SelectedAssociation getModelType()

Get the value of Model.

Returns:

Value of Model.

setModelType

public void setModelType(SelectedAssociation newMethod)

Set the value of Model.

Parameters:

newMethod - Value to assign to Model.

getVerbosity

public int getVerbosity()

Get the value of Verbosity.

Returns:

Value of Verbosity.

setVerbosity

public void setVerbosity(int v)

Set the value of Verbosity.

Parameters:

v - Value to assign to Verbosity.

forName

public static M5 forName(java.lang.String classifierName,
                         java.lang.String[] options)
                  throws java.lang.Exception

Allows to construct the M5 classifier calling it by its name and options.

Parameters:

classifierName - name of the classifier.

options - Options used.

Returns:

Instance of the M5 classifier.

Throws:

java.lang.Exception

hasEnumAttr

public static final boolean hasEnumAttr(M5Instances inst)

Tests if enumerated attribute(s) exists in the instances

Parameters:

inst - instances

Returns:

true if there is at least one; false if none

hasMissing

public static final boolean hasMissing(M5Instances inst)

Tests if missing value(s) exists in the instances

Parameters:

inst - instances

Returns:

true if there is missing value(s); false if none

sum

public static final double sum(int attr,
                               M5Instances inst)

Returns the sum of the instances values of an attribute

Parameters:

attr - an attribute

inst - instances

Returns:

the sum value

sqrSum

public static final double sqrSum(int attr,
                                  M5Instances inst)

Returns the squared sum of the instances values of an attribute

Parameters:

attr - an attribute

inst - instances

Returns:

the squared sum value

stdDev

public static final double stdDev(int attr,
                                  M5Instances inst)

Returns the standard deviation value of the instances values of an attribute

Parameters:

attr - an attribute

inst - instances

Returns:

the standard deviation value

absDev

public static final double absDev(int attr,
                                  M5Instances inst)

Returns the absolute deviation value of the instances values of an attribute

Parameters:

attr - an attribute

inst - instances

Returns:

the absolute deviation value

variance

public static final double variance(int attr,
                                    M5Instances inst)

Returns the variance value of the instances values of an attribute

Parameters:

attr - an attribute

inst - instances

Returns:

the variance value

roundDouble

public static final long roundDouble(double value)

Rounds a double

Parameters:

value - the double value

Returns:

the double rounded

floorDouble

public static final long floorDouble(double value)

Returns the largest (closest to positive infinity) long integer value that is not greater than the argument.

Parameters:

value - the double value

Returns:

the floor integer

doubleToStringF

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

Rounds a double and converts it into a formatted right-justified String. It is like %f format in C language.

Parameters:

value - the double value

width - the width of the string

afterDecimalPoint - the number of digits after the decimal point

Returns:

the double as a formatted string

doubleToStringG

public static final java.lang.String doubleToStringG(double value,
                                                     int width,
                                                     int precision)

Rounds a double and converts it into a formatted right-justified String. If the double is not equal to zero and not in the range [10e-3,10e7] it is returned in scientific format. It is like %g format in C language.

Parameters:

value - the double value

width - the width of the string

precision - the number of valid digits

Returns:

the double as a formatted string

deleteTrailingZerosAndDot

public static final java.lang.StringBuffer deleteTrailingZerosAndDot(java.lang.StringBuffer stringBuffer)

Deletes the trailing zeros and decimal point in a stringBuffer

Parameters:

stringBuffer - string buffer

Returns:

string buffer with deleted trailing zeros and decimal point

smoothenValue

public static final double smoothenValue(double p,
                                         double q,
                                         int n,
                                         int k)

Returns the smoothed values according to the smoothing formula (np+kq)/(n+k)

Parameters:

p - a double, normally is the prediction of the model at the current node

q - a double, normally is the prediction of the model at the up node

n - the number of instances at the up node

k - the smoothing constance, default =15

Returns:

the smoothed value

correlation

public static final double correlation(double[] y1,
                                       double[] y2,
                                       int n)

Returns the correlation coefficient of two double vectors

Parameters:

y1 - double vector 1

y2 - double vector 2

n - the length of two double vectors

Returns:

the correlation coefficient

eqDouble

public static final boolean eqDouble(double a,
                                     double b)

Tests if two double values are equal to each other

Parameters:

a - double 1

b - double 2

Returns:

true if equal; false if not equal

errorMsg

public static final void errorMsg(java.lang.String err)

Prints error message and exits

Parameters:

err - error message

separatorToString

public static final java.lang.String separatorToString()

Prints sepearating line

Returns:

sepearating line string.

headToString

public static final java.lang.String headToString()

Prints the head lines of the output

Returns:

head lines of the output as string.

main

public static void main(java.lang.String[] argv)

Main method for M5' algorithm

Parameters:

argv - command line arguments