Package moa.classifiers.rules.multilabel.attributeclassobservers

Interface AttributeStatisticsObserver

All Superinterfaces:

Configurable, MOAObject, OptionHandler, java.io.Serializable

All Known Subinterfaces:

NominalStatisticsObserver, NumericStatisticsObserver

All Known Implementing Classes:

MultiLabelBSTree, MultiLabelBSTreeFloat, MultiLabelNominalAttributeObserver

public interface AttributeStatisticsObserver
extends OptionHandler

Interface for observing the statistics for an attribute. This observer monitors the statistics for a given attribute. Used in AMRulesMultiLabelLearners

Version:

$Revision: 1 $

Author:

João Duarte ([email protected])

Method Summary

Modifier and Type	Method	Description
AttributeExpansionSuggestion	getBestEvaluatedSplitSuggestion(MultiLabelSplitCriterion criterion, DoubleVector[] preSplitStatistics, int inputAttributeIndex)	Gets the best split suggestion given a criterion and a class distribution
void	observeAttribute(double inputAttributeValue, DoubleVector[] statistics)	Updates statistics of this observer given an attribute value, the index of the statistic and the weight of the instance observed

Methods inherited from interface moa.MOAObject

getDescription, measureByteSize

Methods inherited from interface moa.options.OptionHandler

copy, getCLICreationString, getOptions, getPurposeString, prepareForUse, prepareForUse

Method Detail

observeAttribute

void observeAttribute(double inputAttributeValue,
                      DoubleVector[] statistics)

Updates statistics of this observer given an attribute value, the index of the statistic and the weight of the instance observed

Parameters:

inputAttributeValue - the value for the attribute attribute

statistics - numOutputs x numStatistics the index of the statistic to store

getBestEvaluatedSplitSuggestion

AttributeExpansionSuggestion getBestEvaluatedSplitSuggestion(MultiLabelSplitCriterion criterion,
                                                             DoubleVector[] preSplitStatistics,
                                                             int inputAttributeIndex)

Gets the best split suggestion given a criterion and a class distribution

Parameters:

criterion - the split criterion to use

preSplitStatistics - the class distribution before the split

inputAttributeIndex - the input attribute index

Returns:

suggestion of best attribute split