uk.ac.ed.inf.enzml.weka
Class AttributeFactory

java.lang.Object
  uk.ac.ed.inf.enzml.weka.AttributeFactory

Direct Known Subclasses:

MulanAttributeFactory

public class AttributeFactory
extends java.lang.Object

The attribute factory produces binary and nominal Weka attributes. A binary attribute is a special nominal attribute with only 2 values: yes/no (or present/absent or zero/one). The attribute name often represents the characteristic that can be present or absent. Example: if the attribute name is a protein domain, the value yes/no represents whether the domain is present or absent in that protein sequence (instance). A nominal attribute is an attribute with a set of (more than 2) possible string values.

Version:

20 Feb 2008

Author:

Luna De Ferrari luna.deferrari-at-ed.ac.uk

Field Summary

static java.lang.String	ABSENT_OPTION
static java.lang.String	PRESENT_OPTION

Constructor Summary

AttributeFactory(DataSetGenerator dataSetGenerator)
Takes the attribute names from the dataset generator object.

Method Summary

static boolean	attributeIsValid(weka.core.Attribute attribute) Check if an attribute is valid (not null, non-empty name)
weka.core.Attribute	generateBinaryAttribute(java.lang.String attributeName) Generates a binary attribute
weka.core.Attribute	generateNominalAttribute(java.lang.String attributeName, java.util.Vector<java.lang.String> values) Generate a nominal attribute (with a list of possible values).
weka.core.Attribute	generateNominalAttributeWithDummy(java.lang.String attributeName, java.util.Set<java.lang.String> classValues) Generate a nominal attribute (with a list of possible values).
java.lang.String	getAttributeAbsentOption()
weka.core.Attribute	getAttributeByName(java.lang.String name)
java.lang.String	getAttributePresentOption()
java.util.Vector<weka.core.Attribute>	getAttributes()
java.util.TreeSet<java.lang.String>	getAttributesNames()
java.util.Vector<java.lang.String>	getBinaryAttributeValuesSet() Get the possible attribute values for a yes/no attribute.
weka.core.Attribute	getClassAttribute() Get last attribute as class attribute
DataSetGenerator	getDataSetGenerator()

Methods inherited from class java.lang.Object

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

Field Detail

ABSENT_OPTION

public static final java.lang.String ABSENT_OPTION

See Also:

Constant Field Values

PRESENT_OPTION

public static final java.lang.String PRESENT_OPTION

See Also:

Constant Field Values

Constructor Detail

AttributeFactory

public AttributeFactory(DataSetGenerator dataSetGenerator)

Takes the attribute names from the dataset generator object.

Parameters:

dataSetGenerator -

Method Detail

generateBinaryAttribute

public weka.core.Attribute generateBinaryAttribute(java.lang.String attributeName)

Generates a binary attribute

Parameters:

attributeName - the attribute name

Returns:

the weka attribute

generateNominalAttribute

public weka.core.Attribute generateNominalAttribute(java.lang.String attributeName,
                                                    java.util.Vector<java.lang.String> values)

Generate a nominal attribute (with a list of possible values).

Parameters:

attributeName - the attribute name

Returns:

the weka attribute

generateNominalAttributeWithDummy

public weka.core.Attribute generateNominalAttributeWithDummy(java.lang.String attributeName,
                                                             java.util.Set<java.lang.String> classValues)

Generate a nominal attribute (with a list of possible values). Inserts a dummy value as first element to solve the problem at http://weka.wikispaces.com/ARFF+%28stable+version%29

Parameters:

attributeName - the attribute name

Returns:

the weka attribute

getAttributeAbsentOption

public java.lang.String getAttributeAbsentOption()

getAttributeByName

public weka.core.Attribute getAttributeByName(java.lang.String name)

getAttributePresentOption

public java.lang.String getAttributePresentOption()

getAttributes

public java.util.Vector<weka.core.Attribute> getAttributes()

getAttributesNames

public java.util.TreeSet<java.lang.String> getAttributesNames()

getBinaryAttributeValuesSet

public java.util.Vector<java.lang.String> getBinaryAttributeValuesSet()

Get the possible attribute values for a yes/no attribute. Beware!: the option "no" has to be listed as first option for the sparse ARFF format to work! http://weka.wikispaces.com/ARFF+%28stable+version%29

Returns:

The vector with the two values, for example: ["no", "yes"].

getClassAttribute

public weka.core.Attribute getClassAttribute()

Get last attribute as class attribute

Returns:

a Weka Attribute

getDataSetGenerator

public DataSetGenerator getDataSetGenerator()

attributeIsValid

public static boolean attributeIsValid(weka.core.Attribute attribute)

Check if an attribute is valid (not null, non-empty name)

Parameters:

attribute -

Returns:

true if the attribute is valid (not null, non-empty name), false otherwise