de.uni_leipzig.bf.cluster
Class BorderFlow

java.lang.Object
  de.uni_leipzig.bf.cluster.BorderFlow

public class BorderFlow
extends java.lang.Object

Author:

an

Field Summary

protected ClusterCache	cache
protected java.sql.Connection	con
protected java.lang.String	database
protected ClusterGraph	graph
protected Harden	harden
protected java.lang.String	ip
protected java.lang.String	password
protected java.lang.String	user

Constructor Summary

BorderFlow(java.lang.String file, Harden p_harden)
Constructs new instance for tabulator separated files.

BorderFlow(java.lang.String file, java.lang.String separator, Harden p_harden)
Constructs new instance.

Method Summary

void	closeConnection() Close connection to database.
java.lang.String	cluster(double connThreshold, boolean testOne, boolean heuristic, boolean caching) Starts Borderflow with the given configuration to write in String.
java.lang.String	cluster(java.lang.String[] p_seeds, boolean testOne, boolean heuristic, boolean caching) Starts Borderflow with the given configuration to write in String.
void	clusterToDatabase(double connThreshold) Clusters to the database found under connection
void	clusterToFile(java.lang.String outputFile, double connThreshold, boolean testOne, boolean heuristic, boolean caching) Starts Borderflow with the given configuration to write in file.
double	compareSet(java.util.Set<java.lang.Integer> set1, java.util.Set<java.lang.Integer> set2)
double	getAbsoluteFlow(Cluster c, java.lang.Integer node) Computes the inverse relative flow to make sure that no division by zero is carried out
java.sql.Connection	getConnection() Get connection to database
double	getFlowFromNodeToSet(java.util.TreeSet<java.lang.Integer> nodeSet, java.lang.Integer node) Compute flow from node to set of nodes
double	getInverseRelativeFlow(Cluster c, java.lang.Integer node) Computes the inverse relative flow to make sure that no division by zero is carried out
double	getRelativerFlow(java.util.TreeSet<java.lang.Integer> nodes)
java.lang.String	getSilhouetteAndRelativeFlow(java.util.TreeSet<java.lang.Integer> nodes) Computes the silhouette for a set of nodes
protected java.util.HashMap<java.util.TreeSet<java.lang.Integer>,java.util.TreeSet<java.lang.Integer>>	harden(java.util.HashMap<java.util.TreeSet<java.lang.Integer>,java.util.TreeSet<java.lang.Integer>> clusterSeedMap) Harden method used a Harden instance.
java.util.TreeSet<java.lang.Integer>	heuristicGetCluster(java.lang.Integer seed) Compute cluster for a given seed using heuristic version of BorderFlow.
java.util.TreeSet<java.lang.Integer>	heuristicGetClusterTestOne(java.lang.Integer seed) Compute cluster for a given seed using heuristic and test one version of BorderFlow.
java.util.TreeSet<java.lang.Integer>	knnCluster(java.lang.Integer seed, int k) Compute cluster for a given seed.
void	knnToFile(java.lang.String outputFile, double connThreshold, int k) Starts knn with the given configuration to write in file.
java.lang.String	knnToString(double connThreshold, int k) Starts knn with the given configuration to write in String.
java.lang.String	knnToString(java.lang.String[] p_seeds, int k) Starts knn with the given configuration to write in String.
java.util.TreeSet<java.lang.Integer>	optimalGetCluster(java.lang.Integer seed) Compute cluster for a given seed using optimal version of BorderFlow.
java.util.TreeSet<java.lang.Integer>	optimalGetClusterTestOne(java.lang.Integer seed) Compute cluster for a given seed using optimal and test one version of BorderFlow.
void	setDatabaseVariables(java.lang.String _ip, java.lang.String _database, java.lang.String _user, java.lang.String _password) Set database variables.
protected java.lang.String	writeToString(java.util.HashMap<java.util.TreeSet<java.lang.Integer>,java.util.TreeSet<java.lang.Integer>> clusterSeedMap) Writes clusterSeedMap to String.

Methods inherited from class java.lang.Object

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

Field Detail

cache

protected ClusterCache cache

con

protected java.sql.Connection con

database

protected java.lang.String database

graph

protected ClusterGraph graph

harden

protected Harden harden

ip

protected java.lang.String ip

password

protected java.lang.String password

user

protected java.lang.String user

Constructor Detail

BorderFlow

public BorderFlow(java.lang.String file,
                  Harden p_harden)

Constructs new instance for tabulator separated files.

Parameters:

file - path to file

p_harden - a Harden instance.

BorderFlow

public BorderFlow(java.lang.String file,
                  java.lang.String separator,
                  Harden p_harden)

Constructs new instance.

Parameters:

file - path to file

separator - file separator

p_harden - a Harden instance.

Method Detail

closeConnection

public void closeConnection()

Close connection to database.

cluster

public java.lang.String cluster(double connThreshold,
                                boolean testOne,
                                boolean heuristic,
                                boolean caching)

Starts Borderflow with the given configuration to write in String.

Parameters:

connThreshold - gives the percentage of the maximal connectivity that a node can maximally have to be used as seed

testOne - sets the type of test to use

heuristic - heuristic or the optimal version of borderflow

caching - sets cache version or not

Returns:

clusters

cluster

public java.lang.String cluster(java.lang.String[] p_seeds,
                                boolean testOne,
                                boolean heuristic,
                                boolean caching)

Starts Borderflow with the given configuration to write in String.

Parameters:

p_seeds - Strings to be used as seed

testOne - sets the type of test to use

heuristic - heuristic or the optimal version of borderflow

caching - sets cache version or not

Returns:

clusters

clusterToDatabase

public void clusterToDatabase(double connThreshold)

Clusters to the database found under connection

clusterToFile

public void clusterToFile(java.lang.String outputFile,
                          double connThreshold,
                          boolean testOne,
                          boolean heuristic,
                          boolean caching)

Starts Borderflow with the given configuration to write in file.

Parameters:

outputFile - Output is written in outputFile

connThreshold - gives the percentage of the maximal connectivity that a node can maximally have to be used as seed

testOne - sets the type of test to use

heuristic - heuristic or the optimal version of borderflow

caching - sets cache version or not

compareSet

public double compareSet(java.util.Set<java.lang.Integer> set1,
                         java.util.Set<java.lang.Integer> set2)

getAbsoluteFlow

public double getAbsoluteFlow(Cluster c,
                              java.lang.Integer node)

Computes the inverse relative flow to make sure that no division by zero is carried out

getConnection

public java.sql.Connection getConnection()

Get connection to database

getFlowFromNodeToSet

public double getFlowFromNodeToSet(java.util.TreeSet<java.lang.Integer> nodeSet,
                                   java.lang.Integer node)

Compute flow from node to set of nodes

getInverseRelativeFlow

public double getInverseRelativeFlow(Cluster c,
                                     java.lang.Integer node)

Computes the inverse relative flow to make sure that no division by zero is carried out

getRelativerFlow

public double getRelativerFlow(java.util.TreeSet<java.lang.Integer> nodes)

getSilhouetteAndRelativeFlow

public java.lang.String getSilhouetteAndRelativeFlow(java.util.TreeSet<java.lang.Integer> nodes)

Computes the silhouette for a set of nodes

harden

protected java.util.HashMap<java.util.TreeSet<java.lang.Integer>,java.util.TreeSet<java.lang.Integer>> harden(java.util.HashMap<java.util.TreeSet<java.lang.Integer>,java.util.TreeSet<java.lang.Integer>> clusterSeedMap)

Harden method used a Harden instance.

Parameters:

clusterSeedMap - soft clusters

Returns:

clusterSeedMap hard clusters

heuristicGetCluster

public java.util.TreeSet<java.lang.Integer> heuristicGetCluster(java.lang.Integer seed)

Compute cluster for a given seed using heuristic version of BorderFlow.

heuristicGetClusterTestOne

public java.util.TreeSet<java.lang.Integer> heuristicGetClusterTestOne(java.lang.Integer seed)

Compute cluster for a given seed using heuristic and test one version of BorderFlow.

knnCluster

public java.util.TreeSet<java.lang.Integer> knnCluster(java.lang.Integer seed,
                                                       int k)

Compute cluster for a given seed.

knnToFile

public void knnToFile(java.lang.String outputFile,
                      double connThreshold,
                      int k)

Starts knn with the given configuration to write in file.

knnToString

public java.lang.String knnToString(double connThreshold,
                                    int k)

Starts knn with the given configuration to write in String.

knnToString

public java.lang.String knnToString(java.lang.String[] p_seeds,
                                    int k)

Starts knn with the given configuration to write in String.

optimalGetCluster

public java.util.TreeSet<java.lang.Integer> optimalGetCluster(java.lang.Integer seed)

Compute cluster for a given seed using optimal version of BorderFlow.

optimalGetClusterTestOne

public java.util.TreeSet<java.lang.Integer> optimalGetClusterTestOne(java.lang.Integer seed)

Compute cluster for a given seed using optimal and test one version of BorderFlow.

setDatabaseVariables

public void setDatabaseVariables(java.lang.String _ip,
                                 java.lang.String _database,
                                 java.lang.String _user,
                                 java.lang.String _password)

Set database variables.

writeToString

protected java.lang.String writeToString(java.util.HashMap<java.util.TreeSet<java.lang.Integer>,java.util.TreeSet<java.lang.Integer>> clusterSeedMap)

Writes clusterSeedMap to String.

Parameters:

clusterSeedMap - with clusters and seeds

Returns:

clusters