DynComm (version 2020.1.6)

DynComm: DynComm

Description

Provides a single interface for all algorithms in the different languages.

Usage

DynComm(Algorithm,Criterion,Parameters)

Arguments

Algorithm

One of the available ALGORITHM. Default ALGORITHM$LOUVAIN. See ALGORITHM

Criterion

One of the available CRITERION. Default CRITERION$MODULARITY. See CRITERION

Parameters

A two column matrix defining additional parameters. Default NULL. See the PARAMETERS section on this page

Value

DynComm object

PARAMETERS

A two column matrix defining additional parameters to be passed to the selected ALGORITHM and CRITERION. The first column names the parameter and the second defines its value.

c

Owsinski-Zadrozny quality function parameter. Values [0.0:1.0]. Default: 0.5

k

Shi-Malik quality function kappa_min value. Value > 0 . Default 1

w

Treat graph as weighted. In other words, do not ignore weights for edges that define them when inserting edges in the graph. A weight of exactly zero removes the edge instead of inserting so its weight is never ignored. Without this parameter defined or for edges that do not have a weight defined, edges are assigned the default value of 1 (one). As an example, reading from a file may define weights (a third column) for some edges (defined in rows, one per row) and not for others. With this parameter defined, the edges that have weights that are not exactly zero, have their weight replaced by the default value. Values TRUE,FALSE. Default FALSE

e

Stops when, on a cycle of the algorithm, the quality is increased by less than the value given in this parameter. Value > 0 . Default 0.01

cv

Community-Vertex. Boolean parameter that indicates if sending community mapping to a file prints the community first, if true, or the vertex first, if false. See communityMapping for details. Default TRUE

Methods

postProcess(actions)

Set a list of post processing steps. See postProcess

select(postProcessing,id)

Select between getting the results of the algorithm or one of the post processing steps. See select

results(differential)

Get additional results of the algorithm or the currently selected post processing steps. See results

addRemoveEdges(graphAddRemove)

Add and remove edges read from a matrix or file. See addRemoveEdges

addRemove(graphAddRemove)

Alias for addRemoveEdges(). See addRemoveEdges

add(graphAddRemove)

Alias for addRemoveEdges(). See addRemoveEdges

quality()

Get the quality measurement of the graph after the last iteration. See quality

communityCount()

Get the number of communities after the last iteration. See communityCount

communities()

Get all communities after the last iteration. See communities

communitiesEdgeCount()

Get the number of community to community edges in the graph. See communitiesEdgeCount

communityNeighbours(community)

Get the neighbours of the given community after the last iteration. See communityNeighbours

communityInnerEdgesWeight(community)

Get the sum of weights of the inner edges of the given community after the last iteration. See communityInnerEdgesWeight

communityTotalWeight(community)

Get the sum of weights of all edges of the given community after the last iteration. See communityTotalWeight

communityEdgeWeight(source,destination)

Get the weight of the edge that goes from source community to destination community after the last iteration. See communityEdgeWeight

communityVertexCount(community)

Get the amount of vertices in the given community after the last iteration. See communityVertexCount

communityNodeCount(community)

Alias for communityVertexCount(). See communityVertexCount

community(vertex)

Get the community of the given vertex after the last iteration. See community

vertexCount()

Get the total number of vertices after the last iteration. See vertexCount

nodesCount()

Alias for vertexCount(). See vertexCount

verticesAll()

Get all vertices in the graph after the last iteration. See verticesAll

nodesAll()

Alias for verticesAll(). See verticesAll

neighbours(vertex)

Get the neighbours of the given vertex after the last iteration. See neighbours

edgeWeight(source,destination)

Get the weight of the edge that goes from source vertex to destination vertex after the last iteration. See edgeWeight

edge(source,destination)

Alias for edgeWeight(). See edgeWeight

vertices(community)

Get all vertices belonging to the given community after the last iteration. See vertices

nodes(community)

Alias for vertices(community). See vertices

edgeCount()

Get the number of vertex to vertex edges in the graph. See edgeCount

communityMapping(differential, file)

Get the community mapping for all communities after the last iteration. See communityMapping

time(differential)

Get the cumulative time spent on processing after the last iteration. See time

version()

Get the source code versions of the different sources. See version

Details

Includes methods to get results of processing and to interact with the vertices, edges and communities. Provided methods to return information on the graph are divided into two layers. A lower level layer that interacts with vertices and how they connect. And a higher level layer that interacts with communities and how they connect. Besides the main algorithm, also accepts post processing algorithms that are used mainly to filter the results. Post processing algorithms can use additional computational resources so check the Performance section of the help page of each algorithm you intend to use.

See Also

DynComm-package

Examples

# NOT RUN {
Parameters<-matrix(c("e","0.1","w", "FALSE"),ncol=2, byrow=TRUE)
dc<-DynComm(ALGORITHM$LOUVAIN,CRITERION$MODULARITY,Parameters)
dc$addRemoveEdges(
matrix(
c(10,20,10,30,20,30,30,60,40,60,40,50,50,70,60,70)
,ncol=2,byrow=TRUE)
)
## or
## dc$addRemoveEdges("initial_graph.txt")
dc$communityCount()
## You can use the non inline version of the functions
DynComm.communities(dc)
## Several alias have been defined.
## In this case, communityNodeCount is alias of communityVertexCount
dc$communityNodeCount(10)
dc$communityNeighbours(10)
dc$communityInnerEdgesWeight(10)
dc$communityTotalWeight(10)
dc$communityEdgeWeight(10,40)
dc$community(10) ##this parameter is a vertex not a community. Do not confuse them 
dc$vertices(10)
dc$communityMapping(TRUE)
dc$quality()
dc$time()
## lets add post processing :)
dc$postProcess(
list(
list(POSTPROCESSING$DENSOPT)
)
)
## the results of the last step of post processing are selected automatically
## densopt post processing algorithm may change the community mapping so...
## check it
dc$communityMapping(TRUE)
## densopt post processing algorithm may change quality so check it
dc$quality()
## time is now the total time of the main algorithm plus the time of every...
## post processing algorithm up to the one selected
dc$time()
## get back to main algorithm results to check they haven't changed
dc$select(POSTPROCESSING$NONE)
dc$communityMapping(TRUE)
dc$quality()
dc$time()
## add and remove edges. Notice that there is one more column to give...
## weights of zero on the edges to remove. In this case, all other weights...
## are ignored because the graph is set to ignore weights (parameter w is...
## false).
dc$addRemoveEdges(
matrix(
c(30,60,0,40,60,0.23,10,80,2342,80,90,3.1415)
,ncol=3,byrow=TRUE)
)
## since the post processing was not reset, it will be automatically...
## calculated and results switched to the last step. In this case, to the...
## densopt algorithm
dc$communityMapping(TRUE)
dc$quality()
dc$time()
## get back to main algorithm results to check them
dc$select(POSTPROCESSING$NONE)
dc$communityMapping(TRUE)
dc$quality()
dc$time()
## lets reset/remove post processing
dc$postProcess()


# }