de.lmu.ifi.dbs.elki.datasource.filter
Class AbstractRandomFeatureSelectionFilter<V extends FeatureVector<?,?>>

java.lang.Object
  de.lmu.ifi.dbs.elki.datasource.filter.AbstractConversionFilter<V,V>
      de.lmu.ifi.dbs.elki.datasource.filter.AbstractRandomFeatureSelectionFilter<V>

Type Parameters:

V - the type of FeatureVector contained in both the original data of the base parser and the projected data of this ProjectionParser

All Implemented Interfaces:

ObjectFilter

Direct Known Subclasses:

DoubleVectorRandomProjectionFilter, SparseFloatVectorRandomProjectionFilter

public abstract class AbstractRandomFeatureSelectionFilter<V extends FeatureVector<?,?>>
extends AbstractConversionFilter<V,V>

A RandomProjectionParser selects a subset of attributes randomly for projection of a ParsingResult.

The cardinality of the subset of attributes is specified as a parameter.

Nested Class Summary

static class

AbstractRandomFeatureSelectionFilter.Parameterizer<V extends NumberVector<V,?>> Parameterization class.

Field Summary

protected int	k Holds the desired cardinality of the subset of attributes selected for projection.
static OptionID	NUMBER_SELECTED_ATTRIBUTES_ID Parameter for the desired cardinality of the subset of attributes selected for projection.
protected Random	random Holds a random object.
protected BitSet	selectedAttributes The selected attributes

Constructor Summary

AbstractRandomFeatureSelectionFilter(int dim)
Constructor.

Method Summary

protected boolean

prepareStart(SimpleTypeInformation<V> in) Return "true" when the normalization needs initialization (two-pass filtering!)

Methods inherited from class de.lmu.ifi.dbs.elki.datasource.filter.AbstractConversionFilter

convertedType, filter, filterSingleObject, getInputTypeRestriction, prepareComplete, prepareProcessInstance

Methods inherited from class java.lang.Object

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

Field Detail

selectedAttributes

protected BitSet selectedAttributes

The selected attributes

NUMBER_SELECTED_ATTRIBUTES_ID

public static final OptionID NUMBER_SELECTED_ATTRIBUTES_ID

Parameter for the desired cardinality of the subset of attributes selected for projection.

Key: -randomprojection.numberselected

Default: 1

Constraint: ≥1

k

protected int k

Holds the desired cardinality of the subset of attributes selected for projection.

random

protected final Random random

Holds a random object.

Constructor Detail

AbstractRandomFeatureSelectionFilter

public AbstractRandomFeatureSelectionFilter(int dim)

Constructor.

Parameters:

dim - dimensionality

Method Detail

prepareStart

protected boolean prepareStart(SimpleTypeInformation<V> in)

Description copied from class: AbstractConversionFilter

Return "true" when the normalization needs initialization (two-pass filtering!)

Overrides:

prepareStart in class AbstractConversionFilter<V extends FeatureVector<?,?>,V extends FeatureVector<?,?>>

Parameters:

in - Input type information

Returns:

true or false