PageRank Class Reference

PageRank class. More...

#include <PageRank.h>

Public Member Functions
	PageRank (sparksee::gdb::Session &session)
	Builds the PageRank.
virtual	~PageRank ()
	Destructor.
void	AddEdgeType (sparksee::gdb::type_t type, sparksee::gdb::EdgesDirection dir) throw (sparksee::gdb::Error)
	Allows for traversing edges of the given type.
void	AddAllEdgeTypes (sparksee::gdb::EdgesDirection dir)
	Allows for traversing all edge types of the graph.
void	AddNodeType (sparksee::gdb::type_t type) throw (sparksee::gdb::Error)
	Allows for traversing nodes of the given type.
void	AddAllNodeTypes ()
	Allows for traversing all node types of the graph.
void	SetNumIterations (sparksee::gdb::int32_t numIterations)
	Sets the number of iterations to run the PageRank for.
void	SetTolerance (sparksee::gdb::double64_t tolerance)
	Sets the tolerance threashold to continue computing the PageRank after each iteration.
void	SetDamping (sparksee::gdb::double64_t damping)
	Sets the damping value for the PageRank.
void	SetInitialPageRankValue (sparksee::gdb::double64_t startValue)
	Sets the initial PageRank value.
void	SetOutputAttributeType (sparksee::gdb::attr_t attr) throw (sparksee::gdb::Error)
	Sets the output attribute type.
void	SetEdgeWeightAttributeType (sparksee::gdb::attr_t attr) throw (sparksee::gdb::Error)
	Sets the attribute to use as edge weight.
void	SetDefaultWeight (sparksee::gdb::double64_t weight)
	Sets the default weight for those cases when a given edge does not have a weight attribute set.
void	SetStartingNode (sparksee::gdb::oid_t startNode) throw (sparksee::gdb::Error)
	Sets the starting node of the page rank to compute the Personalized PageRank variant.
void	Run () throw (sparksee::gdb::Error)
	Runs the algorithm.

Detailed Description

PageRank class.

Implements the PageRank algorithm

Author

Sparsity Technologies http://www.sparsity-technologies.com

Constructor & Destructor Documentation

PageRank()

PageRank::PageRank

(

sparksee::gdb::Session &

session

)

Builds the PageRank.

Parameters

session

[in] The session to use

Member Function Documentation

AddAllEdgeTypes()

void PageRank::AddAllEdgeTypes

(

sparksee::gdb::EdgesDirection

dir

)

Allows for traversing all edge types of the graph.

The direction is interpreted as in which direction an edge can be followed from a node to influence other nodes.

Parameters

dir	[in] Edge direction.

AddEdgeType()

void PageRank::AddEdgeType	(	sparksee::gdb::type_t	type,
		sparksee::gdb::EdgesDirection	dir )
throw	(	sparksee::gdb::Error )

Allows for traversing edges of the given type.

If the edge type was already added, the existing direction is overwritten The direction is interpreted as in which direction an edge can be followed from a node to influence other nodes.

Parameters

type	[in] Edge type.
dir	[in] Edge direction.

Run()

void PageRank::Run	(		)
throw	(	sparksee::gdb::Error )

Runs the algorithm.

Exceptions

sparksee::gdb::Error

SetDamping()

void PageRank::SetDamping

(

sparksee::gdb::double64_t

damping

)

Sets the damping value for the PageRank.

Parameters

damping

[in] The damping value. Default: 0.85

SetDefaultWeight()

void PageRank::SetDefaultWeight

(

sparksee::gdb::double64_t

weight

)

Sets the default weight for those cases when a given edge does not have a weight attribute set.

Default: 0.0

Parameters

weight

[in] The default weight

SetEdgeWeightAttributeType()

void PageRank::SetEdgeWeightAttributeType	(	sparksee::gdb::attr_t	attr	)
throw	(	sparksee::gdb::Error )

Sets the attribute to use as edge weight.

If the multiple edge are set for traversal, this attribute must be of type GLOBAL_TYPE or EDGES_TYPE. Additionally, the attribute must be of type Double. Finally, negative weights are treated as non existing, so the default weight applies.

Parameters

attr	[in] The attribute type to use as a weight. Default: InvalidAttribute

Exceptions

sparksee::gdb::Error

SetInitialPageRankValue()

void PageRank::SetInitialPageRankValue

(

sparksee::gdb::double64_t

startValue

)

Sets the initial PageRank value.

If a starting node is set, this initial value is only set for the starting node and the rest of nodes are set to 0.0

Parameters

startValue

[in] The initial value to set. Default: 0.0

SetNumIterations()

void PageRank::SetNumIterations

(

sparksee::gdb::int32_t

numIterations

)

Sets the number of iterations to run the PageRank for.

Parameters

numIterations

[in] The number of interations to set. Default: 20

SetOutputAttributeType()

void PageRank::SetOutputAttributeType	(	sparksee::gdb::attr_t	attr	)
throw	(	sparksee::gdb::Error )

Sets the output attribute type.

If the PageRank will run on more than one node type, then the output attribute must be of type GLOBAL_TYPE or NODES_TYPE. Otherwise, it must be a valid attribute for the used node type.

Parameters

attr	[in] The attribute to store the result. Default: InvalidAttribute

SetStartingNode()

void PageRank::SetStartingNode	(	sparksee::gdb::oid_t	startNode	)
throw	(	sparksee::gdb::Error )

Sets the starting node of the page rank to compute the Personalized PageRank variant.

Exceptions

sparksee::gdb::Error

SetTolerance()

void PageRank::SetTolerance

(

sparksee::gdb::double64_t

tolerance

)

Sets the tolerance threashold to continue computing the PageRank after each iteration.

If all the changes to any PPR value after an iteration are below that tolerance threshold, the algorithm finishes.

Parameters

tolerance

[in] The tolerance to use normalized between 0 and 1. Default: 0.000001

---

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

• PageRank.h