mulan.dimensionalityReduction
Class Ranker

java.lang.Object
  mulan.dimensionalityReduction.Ranker

public class Ranker
extends Object

Ranks attributes according to an AttributeEvaluator. It internally uses Weka's Ranker, initialized so as to neglect the labels.

Version:

2012.05.30

Author:

Grigorios Tsoumakas

Constructor Summary

Ranker()

Method Summary

int[]

search(AttributeEvaluator attributeEval, MultiLabelInstances mlData) Calls a specified AttributeEvaluator to evaluate each feature attribute of specified MultiLabelInstances data set, excluding labels.

Methods inherited from class java.lang.Object

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

Constructor Detail

Ranker

public Ranker()

Method Detail

search

public int[] search(AttributeEvaluator attributeEval,
                    MultiLabelInstances mlData)
             throws Exception

Calls a specified AttributeEvaluator to evaluate each feature attribute of specified MultiLabelInstances data set, excluding labels. Internally it uses Ranker

Parameters:

attributeEval - the attribute evaluator to guide the search

mlData - the multi-label instances data set

Returns:

an array (not necessarily ordered) of selected attribute indexes

Throws:

Exception - if an error occur in search