uk.ac.ed.inf.csb.taweka.mulan
Class MulanDataSetWriter
java.lang.Object
uk.ac.ed.inf.csb.taweka.DataSetWriter
uk.ac.ed.inf.csb.taweka.mulan.MulanDataSetWriter
public class MulanDataSetWriter
- extends DataSetWriter
This class generates a data set (instances, attributes, classes) in Mulan
style (for multi-label machine learning). The main difference compared with
the classic Weka ARFF, is that there are multiple classes and each is simply
listed as a normal binary {0, 1} attribute. The class attributes must also be
specified in a separate Mulan XML.
- Version:
- 7 May 2010
- Author:
- Luna De Ferrari luna.deferrari-at-ed.ac.uk
|
Constructor Summary |
MulanDataSetWriter(ArffGenerator arffGenerator,
uk.ac.ed.inf.csb.utils.OneToManyMap instanceAttributeMap,
uk.ac.ed.inf.csb.utils.OneToManyMap instanceClassMap)
The set of class attributes. |
| Methods inherited from class uk.ac.ed.inf.csb.taweka.DataSetWriter |
generateDataSet, getArffProperties, getAttributeFactory, getAttributes, getClassName, getDataSetName, getInstanceAttributeData, getInstanceClassData, getInstances, getLog, getProperties, writeInstancesToFile |
| Methods inherited from class java.lang.Object |
equals, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait |
MulanDataSetWriter
public MulanDataSetWriter(ArffGenerator arffGenerator,
uk.ac.ed.inf.csb.utils.OneToManyMap instanceAttributeMap,
uk.ac.ed.inf.csb.utils.OneToManyMap instanceClassMap)
- The set of class attributes. Mulan multi-label learning can accept
multiple class attributes formatted as normal attributes. The Mulan XML
file contains the attributes to be used as class attributes.