---

Cluster Class Reference

Cluster representation, a vector of cluster elements. More...

#include <Cluster.hpp>

Inheritance diagram for Cluster:

List of all members.

Public Methods
	Cluster (int cid, const Index &ind, const SimilarityMethod &sim)
	create empty cluster with specific id.
virtual	~Cluster ()
	delete cluster
void	setId (int cid)
	set cluster id.
void	setName (const string &newName)
	set cluster symbolic name.
int	getId () const
	get cluster id.
const string &	getName () const
	get cluster symbolic name.
const vector< ClusterElt > *	getIds () const
	get vector of elements in this cluster.
vector< int >	getDocIds () const
	get vector of doc ids in this cluster.
virtual ClusterRep *	getClusterRep () const=0
	Get the ClusterRep for this Cluster for scoring. The Rep will have been weighted by the SimilarityMethod. Caller responsible for deleting.
virtual double	score (const ClusterRep *rep) const=0
	score a document against this cluster, given the rep.
virtual void	add (const ClusterElt &elt)
	add the element to this cluster
virtual void	add (const vector< int > docids)
	Add a list of documents to a cluster.
virtual void	remove (const ClusterElt &elt)
	remove the element from this cluster
virtual vector< Cluster * >	split (int numParts=2)
	Split this cluster into subclusters.
virtual void	merge (const Cluster *c)
	merge in a cluster
virtual bool	read (ifstream &in)
	read a cluster in from the cluster db file.
virtual void	write (ofstream &out)
	write a cluster to the cluster db file.
void	print () const
	pretty print to stdout.
int	getSize () const
	How many elements.
virtual double	sum2 () const
	Sum of squared values in the cluster's term vector.
virtual string	getKeyWords (int numTerms=10) const
	Return the top numTerms terms in the cluster (tf.idf scores).
Protected Attributes
int	id
	internal id
string	name
	external symbolic name.
vector< ClusterElt >	ids
	vector of cluster elements in this cluster.
const SimilarityMethod &	similarity
int	size
double	weight
const Index &	ind
fstream *	datFile
long	offset

---

Detailed Description

Cluster representation, a vector of cluster elements.

---

Constructor & Destructor Documentation

Cluster::Cluster (  int    cid, const Index &    ind, const SimilarityMethod &    sim )

create empty cluster with specific id.

Cluster::~Cluster (    )  [virtual]

delete cluster

---

Member Function Documentation

void Cluster::add (  const vector< int >    docids )  [virtual]

Add a list of documents to a cluster. Reimplemented in CentroidCluster.

void Cluster::add (  const ClusterElt &    elt )  [virtual]

add the element to this cluster Reimplemented in CentroidCluster.

virtual ClusterRep* Cluster::getClusterRep (    )  [pure virtual]

Get the ClusterRep for this Cluster for scoring. The Rep will have been weighted by the SimilarityMethod. Caller responsible for deleting. Implemented in AgglomCluster, and CentroidCluster.

vector< int > Cluster::getDocIds (    )

get vector of doc ids in this cluster.

int Cluster::getId (    )

get cluster id.

const vector< ClusterElt > * Cluster::getIds (    )

get vector of elements in this cluster.

string Cluster::getKeyWords (  int    numTerms = 10 )  const [virtual]

Return the top numTerms terms in the cluster (tf.idf scores).

const string & Cluster::getName (    )

get cluster symbolic name.

int Cluster::getSize (    )  const [inline]

How many elements.

void Cluster::merge (  const Cluster *    c )  [virtual]

merge in a cluster

void Cluster::print (    )

pretty print to stdout.

bool Cluster::read (  ifstream &    in )  [virtual]

read a cluster in from the cluster db file. Reimplemented in CentroidCluster.

void Cluster::remove (  const ClusterElt &    elt )  [virtual]

remove the element from this cluster Reimplemented in CentroidCluster.

virtual double Cluster::score (  const ClusterRep *    rep )  const [pure virtual]

score a document against this cluster, given the rep. Implemented in AgglomCluster, and CentroidCluster.

void Cluster::setId (  int    cid )

set cluster id.

void Cluster::setName (  const string &    newName )

set cluster symbolic name.

vector< Cluster * > Cluster::split (  int    numParts = 2 )  [virtual]

Split this cluster into subclusters.

virtual double Cluster::sum2 (    )  const [inline, virtual]

Sum of squared values in the cluster's term vector. Reimplemented in AgglomCluster, and CentroidCluster.

void Cluster::write (  ofstream &    out )  [virtual]

write a cluster to the cluster db file.

---

Member Data Documentation

fstream* Cluster::datFile [protected]

int Cluster::id [protected]

internal id

vector<ClusterElt> Cluster::ids [protected]

vector of cluster elements in this cluster.

const Index& Cluster::ind [protected]

string Cluster::name [protected]

external symbolic name.

long Cluster::offset [protected]

const SimilarityMethod& Cluster::similarity [protected]

int Cluster::size [protected]

double Cluster::weight [protected]

---

The documentation for this class was generated from the following files:

• Cluster.hpp
• cluster/src/Cluster.cpp

---