de.lmu.ifi.dbs.elki.algorithm.clustering.biclustering

Class AbstractBiclustering<V extends NumberVector,M extends BiclusterModel>

java.lang.Object

de.lmu.ifi.dbs.elki.algorithm.AbstractAlgorithm<Clustering<M>>

de.lmu.ifi.dbs.elki.algorithm.clustering.biclustering.AbstractBiclustering<V,M>

Type Parameters:

V - a certain subtype of NumberVector - the data matrix is supposed to consist of rows where each row relates to an object of type V and the columns relate to the attribute values of these objects

M - Cluster model type

All Implemented Interfaces:

Algorithm, ClusteringAlgorithm<Clustering<M>>

Direct Known Subclasses:

ChengAndChurch

public abstract class AbstractBiclustering<V extends NumberVector,M extends BiclusterModel>
extends AbstractAlgorithm<Clustering<M>>
implements ClusteringAlgorithm<Clustering<M>>

Abstract class as a convenience for different biclustering approaches.

The typically required values describing submatrices are computed using the corresponding values within a database of NumberVectors.

The database is supposed to present a data matrix with a row representing an entry (NumberVector), a column representing a dimension (attribute) of the NumberVectors.

Since:

0.2

Author:

Arthur Zimek

Field Summary

Fields

Modifier and Type	Field and Description
private int	colDim Column dimensionality.
private Database	database Keeps the currently set database.
private DBIDArrayIter	iter Iterator to use for more efficient random access.
protected Relation<V>	relation Relation we use.
protected ArrayDBIDs	rowIDs The row ids corresponding to the currently set relation.

Constructor Summary

Constructors

Modifier	Constructor and Description
protected	AbstractBiclustering() Constructor.

Method Summary

Methods

Modifier and Type	Method and Description
protected abstract Clustering<M>	biclustering() Run the actual biclustering algorithm.
protected int[]	colsBitsetToIDs(BitSet cols) Convert a bitset into integer column ids.
protected int[]	colsBitsetToIDs(long[] cols) Convert a bitset into integer column ids.
protected Cluster<BiclusterModel>	defineBicluster(BitSet rows, BitSet cols) Defines a Bicluster as given by the included rows and columns.
protected Cluster<BiclusterModel>	defineBicluster(long[] rows, long[] cols) Defines a Bicluster as given by the included rows and columns.
protected int	getColDim() The number of columns of the data matrix.
Database	getDatabase() Getter for database.
Relation<V>	getRelation() Getter for the relation.
protected DBID	getRowDBID(int row) Deprecated. Expensive!
protected int	getRowDim() The number of rows of the data matrix.
protected ArrayDBIDs	rowsBitsetToIDs(BitSet rows) Convert a bitset into integer row ids.
protected ArrayDBIDs	rowsBitsetToIDs(long[] rows) Convert a bitset into integer row ids.
Clustering<M>	run(Relation<V> relation) Prepares the algorithm for running on a specific database.
protected double	valueAt(int row, int col) Returns the value of the data matrix at row row and column col.

Methods inherited from class de.lmu.ifi.dbs.elki.algorithm.AbstractAlgorithm

getInputTypeRestriction, getLogger, makeParameterDistanceFunction, run

Methods inherited from class java.lang.Object

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

Methods inherited from interface de.lmu.ifi.dbs.elki.algorithm.clustering.ClusteringAlgorithm

run

Methods inherited from interface de.lmu.ifi.dbs.elki.algorithm.Algorithm

getInputTypeRestriction

Field Detail

database

private Database database

Keeps the currently set database.

relation

protected Relation<V extends NumberVector> relation

Relation we use.

iter

private DBIDArrayIter iter

Iterator to use for more efficient random access.

rowIDs

protected ArrayDBIDs rowIDs

The row ids corresponding to the currently set relation.

colDim

private int colDim

Column dimensionality.

Constructor Detail

AbstractBiclustering

protected AbstractBiclustering()

Constructor.

Method Detail

run

public final Clustering<M> run(Relation<V> relation)

Prepares the algorithm for running on a specific database.

Assigns the database, the row ids, and the col ids, then calls biclustering().

Any concrete algorithm should be implemented within method biclustering() by an inheriting biclustering approach.

Parameters:

relation - Relation to process

Returns:

Clustering result

biclustering

protected abstract Clustering<M> biclustering()

Run the actual biclustering algorithm.

This method is supposed to be called only from the method run(de.lmu.ifi.dbs.elki.database.relation.Relation<V>).

colsBitsetToIDs

protected int[] colsBitsetToIDs(BitSet cols)

Convert a bitset into integer column ids.

Parameters:

cols -

Returns:

integer column ids

rowsBitsetToIDs

protected ArrayDBIDs rowsBitsetToIDs(BitSet rows)

Convert a bitset into integer row ids.

Parameters:

rows -

Returns:

integer row ids

defineBicluster

protected Cluster<BiclusterModel> defineBicluster(BitSet rows,
                                      BitSet cols)

Defines a Bicluster as given by the included rows and columns.

Parameters:

rows - the rows included in the Bicluster

cols - the columns included in the Bicluster

Returns:

a Bicluster as given by the included rows and columns

defineBicluster

protected Cluster<BiclusterModel> defineBicluster(long[] rows,
                                      long[] cols)

Defines a Bicluster as given by the included rows and columns.

Parameters:

rows - the rows included in the Bicluster

cols - the columns included in the Bicluster

Returns:

A Bicluster as given by the included rows and columns

valueAt

protected double valueAt(int row,
             int col)

Returns the value of the data matrix at row row and column col.

Parameters:

row - the row in the data matrix according to the current order of rows (refers to database entry database.get(rowIDs[row]))

col - the column in the data matrix according to the current order of rows (refers to the attribute value of an database entry getValue(colIDs[col]))

Returns:

the attribute value of the database entry as retrieved by database.get(rowIDs[row]).getValue(colIDs[col])

getRowDBID

@Deprecated
protected DBID getRowDBID(int row)

Deprecated. Expensive!

Get the DBID of a certain row

Parameters:

row - Row number

Returns:

DBID of this row

colsBitsetToIDs

protected int[] colsBitsetToIDs(long[] cols)

Convert a bitset into integer column ids.

Parameters:

cols -

Returns:

integer column ids

rowsBitsetToIDs

protected ArrayDBIDs rowsBitsetToIDs(long[] rows)

Convert a bitset into integer row ids.

Parameters:

rows -

Returns:

integer row ids

getRowDim

protected int getRowDim()

The number of rows of the data matrix.

Returns:

the number of rows of the data matrix

getColDim

protected int getColDim()

The number of columns of the data matrix.

Returns:

the number of columns of the data matrix

getDatabase

public Database getDatabase()

Getter for database.

Returns:

database

getRelation

public Relation<V> getRelation()

Getter for the relation.

Returns:

relation