net.sf.myra.datamining
Class Classifier

java.lang.Object
  extended by net.sf.myra.datamining.Classifier
Direct Known Subclasses:
SeCo

public abstract class Classifier
extends java.lang.Object

This class defines a skeletal implementation of a classifier algorithm. A classifier instance is responsible for creating a Model instance during training, which in turn can be used to classify unseen instances.

Version:
$Revision$ $Date:: $
Author:
Fernando Esteban Barril Otero
See Also:
Model

Field Summary
static int DEFAULT_MAXIMUM_UNCOVERED_CASES
          The default maximum number of uncovered cases.
static java.lang.String MAXIMUM_UNCOVERED_CASES
          The property key under which the maximum uncovered value is stored.
static java.lang.String RULE_QUALITY_FUNCTION
          The property key under which the rule quality function value is stored.
 
Constructor Summary
Classifier(java.lang.String optionsFile)
          Default constructor.
 
Method Summary
 java.util.Properties filter(java.util.Properties properties)
          Filters the specified properties.
 java.util.Properties getOptions()
          Returns the configuration property options.
abstract  Model run(Dataset dataset)
          Trains the classifier using the specified dataset.
 
Methods inherited from class java.lang.Object
clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait
 

Field Detail

MAXIMUM_UNCOVERED_CASES

public static final java.lang.String MAXIMUM_UNCOVERED_CASES
The property key under which the maximum uncovered value is stored.

See Also:
Constant Field Values

DEFAULT_MAXIMUM_UNCOVERED_CASES

public static final int DEFAULT_MAXIMUM_UNCOVERED_CASES
The default maximum number of uncovered cases.

See Also:
Constant Field Values

RULE_QUALITY_FUNCTION

public static final java.lang.String RULE_QUALITY_FUNCTION
The property key under which the rule quality function value is stored.

See Also:
Constant Field Values
Constructor Detail

Classifier

public Classifier(java.lang.String optionsFile)
Default constructor.

Parameters:
optionsFile - the classifier options filename.
Method Detail

getOptions

public java.util.Properties getOptions()
Returns the configuration property options.

Returns:
the configuration property options.

filter

public java.util.Properties filter(java.util.Properties properties)
Filters the specified properties. The classifier implementation should remove all properties from the framework that are not used.

Parameters:
properties - the framework (default) properties.
Returns:
the filtered properties.

run

public abstract Model run(Dataset dataset)
Trains the classifier using the specified dataset.

Parameters:
dataset - the training dataset.
Returns:
the classification model.
See Also:
Model


Copyright © 2013. All Rights Reserved.