Main Page | Namespace List | Class Hierarchy | Class List | File List | Namespace Members | Class Members | File Members | Related Pages

lemur::cluster::OfflineCluster Class Reference

Offline clustering algorithms. More...

#include <OfflineCluster.hpp>

List of all members.

Public Member Functions

 OfflineCluster (const lemur::api::Index &ind, enum ClusterParam::simTypes simType=ClusterParam::COS, enum ClusterParam::clusterTypes clusterType=ClusterParam::CENTROID, enum ClusterParam::docModes docMode=ClusterParam::DMAX)
 initialize the cluster methods

 ~OfflineCluster ()
 clean up

vector< Cluster * > * kMeans (vector< lemur::api::DOCID_T > docIds, int numParts=2, int maxIters=100)
vector< Cluster * > * kMeans (Cluster *cluster, int numParts=2, int maxIters=100)
 k-means caller responsible for deleting contents of return vector.

vector< Cluster * > * bisecting_kMeans (vector< lemur::api::DOCID_T > docIds, int numParts=2, int numIters=5, int maxIters=100)

Private Member Functions

bool compareClusterSets (Cluster **, Cluster **, int n)
 Are two sets of clusters equal?

vector< lemur::api::DOCID_T > selectSeeds (vector< lemur::api::DOCID_T > docIds, int num)
 Choose num seeds randomly from docIds.

Cluster * chooseSplit (vector< Cluster * > *working)
 Choose largest cluster from working to split.

double scoreSet (vector< Cluster * > *working)
 Score sum of within cluster similarity over a set of clusters.


Private Attributes

const SimilarityMethod * sim
 Similarity Method to use.

ClusterFactory * factory
 Cluster factory.

const lemur::api::Index & index
 Database containing the collection to operate on.

Detailed Description

Offline clustering algorithms.

Constructor & Destructor Documentation

lemur::cluster::OfflineCluster::OfflineCluster (  const lemur::api::Index &  ind,
enum ClusterParam::simTypes  simType = ClusterParam::COS,
enum ClusterParam::clusterTypes  clusterType = ClusterParam::CENTROID,
enum ClusterParam::docModes  docMode = ClusterParam::DMAX
) 
 

initialize the cluster methods

lemur::cluster::OfflineCluster::~OfflineCluster (   ) 
 

clean up

Member Function Documentation

vector< lemur::cluster::Cluster * > * lemur::cluster::OfflineCluster::bisecting_kMeans (  vector< lemur::api::DOCID_T >  docIds,
int  numParts = 2,
int  numIters = 5,
int  maxIters = 100
) 
 

bisecting k-means caller responsible for deleting contents of return vector.

lemur::cluster::Cluster * lemur::cluster::OfflineCluster::chooseSplit (  vector< Cluster * > *  working  )  [private]
 

Choose largest cluster from working to split.

bool lemur::cluster::OfflineCluster::compareClusterSets (  Cluster **  ,
Cluster **  ,
int  n
)  [private]
 

Are two sets of clusters equal?

vector< lemur::cluster::Cluster * > * lemur::cluster::OfflineCluster::kMeans (  Cluster *  cluster,
int  numParts = 2,
int  maxIters = 100
) 
 

k-means caller responsible for deleting contents of return vector.

vector< lemur::cluster::Cluster * > * lemur::cluster::OfflineCluster::kMeans (  vector< lemur::api::DOCID_T >  docIds,
int  numParts = 2,
int  maxIters = 100
) 
 

Cluster a set of documents into numParts partitions (default 2). k-means caller responsible for deleting contents of return vector.

double lemur::cluster::OfflineCluster::scoreSet (  vector< Cluster * > *  working  )  [private]
 

Score sum of within cluster similarity over a set of clusters.

vector< lemur::api::DOCID_T > lemur::cluster::OfflineCluster::selectSeeds (  vector< lemur::api::DOCID_T >  docIds,
int  num
)  [private]
 

Choose num seeds randomly from docIds.

Member Data Documentation

ClusterFactory* lemur::cluster::OfflineCluster::factory [private]
 

Cluster factory.

const lemur::api::Index& lemur::cluster::OfflineCluster::index [private]
 

Database containing the collection to operate on.

const SimilarityMethod* lemur::cluster::OfflineCluster::sim [private]
 

Similarity Method to use.

The documentation for this class was generated from the following files:
Generated on Tue Jun 15 11:03:05 2010 for Lemur by doxygen 1.3.4