We're sorry but this page doesn't work properly without JavaScript enabled. Please enable it to continue.
Feedback

Community detection by graph Voronoi diagrams

00:00

Formal Metadata

Title
Community detection by graph Voronoi diagrams
Title of Series
Number of Parts
49
Author
License
CC Attribution 3.0 Unported:
You are free to use, adapt and copy, distribute and transmit the work or content in adapted or unchanged form for any legal purpose as long as the work is attributed to the author in the manner specified by the author or licensor.
Identifiers
Publisher
Release Date
Language

Content Metadata

Subject Area
Genre
Abstract
Accurate and efficient community detection in networks is a key challenge for complex network theory and its applications. The problem is analogous to cluster analysis in data mining, a field rich in metric space-based methods. Common to these methods is a geometric, distance-based definition of clusters or communities. Here we propose a new geometric approach to graph community detection based on graph Voronoi diagrams. Our method serves as proof of principle that the definition of appropriate distance metrics on graphs can bring a rich set of metric space-based clustering methods to network science. We employ a simple edge metric that reflects the intra- or inter-community character of edges, and a graph density-based rule to identify seed nodes of Voronoi cells. Our algorithm outperforms most network community detection methods applicable to large networks on benchmark as well as real-world networks. In addition to offering a computationally efficient alternative for community detection, our method opens new avenues for adapting a wide range of data mining algorithms to complex networks from the class of centroid- and density-based clustering methods.
Electric power distributionSensorPlain bearingParticle physicsVideoComputer animation
Sensor
KopfstützeSpeckle imagingDrehmasseGroup delay and phase delayCrystal structureTransportPattern (sewing)Lunar nodeSensorDiagram
Cell (biology)TypesettingPartition of a set
TypesettingSEEDElectric generatorCell (biology)Diagram
RulerPartition of a setTypesettingPhotodissoziationElectric generator
Lunar nodeCosmic distance ladderLunar node
Lunar node
RulerEGPRSLunar nodeCell (biology)Electric generatorRulerComputer animation
Electronic mediaPlane (tool)Group delay and phase delayShip classDensityBlackTypesettingDiagram
Diagram
Black holeGreen politicsCell (biology)SEEDDensityComputer animation
YearSEEDMeasurementPlane (tool)Partition of a setComputer animationDiagram
Lunar nodeLastLunar nodeGenerationCartridge (firearms)Speise <Technik>SensorCosmic distance ladderKette <Zugmittel>Electric generatorDensityScale (map)MeasurementDiagram
EGPRSPagerElectric generatorLunar nodeCosmic distance ladderComputer animation
GeokoronaCylinder blockGeokorona
SensorBand gapQuality (business)Computer animation
Computer animation
Computer animation
Transcript: English(auto-generated)
We present a novel community detection algorithm using graph Voronoi diagrams. Many complex systems can be represented as networks. Social networks, transportation networks, networks defined on biological systems are just a few examples. A common challenge in studying complex networks is the identification of community structure.
In general, communities are defined as groups of nodes which are more densely connected to each other than with the rest of the network. However, this is a qualitative description and no common mathematical definition has been agreed upon. There are many different approaches. Clustering problems similar to graph community detection also occur in
data mining, pattern recognition, machine learning and statistical data analysis. These problems, however, are defined on continuous metric spaces simplifying their formulation. Voronoi diagrams are a common way of partitioning metric spaces into regions or cells around a given set of points
called seeds or generator points such that each point of the space belongs to the cell of the closest seed. A similar partitioning can be applied on graphs if we set a few generator nodes, associate positive values to edges and define the distance between any pair of nodes as the shortest path connecting the two nodes.
We have shown that graph Voronoi diagrams can be used to detect communities. We need rules for setting the lengths of edges and choosing the generator nodes. Then communities can be identified as the graph Voronoi cells. To explain the basic concept of our algorithm, we illustrated on a clustering problem defined in Euclidean space.
Imagine a set of points thrown onto a two-dimensional plane, unevenly distributed to form local groups. The question is, how can we partition the space to separate these clusters? First, we define a square lattice across the plane and calculate the local density in each small square as the number of black dots included in it.
Green stands for low, red for high density values. Second, we choose squares with the largest density within the neighborhood of a given radius r as Voronoi cell seeds. In this example, choosing the radius to be of four units yields seven seeds and corresponding neighborhoods.
Third, we assign each point of the plane to the seed closest to it. This yields a Voronoi partitioning diagram. Similar principles can be applied to obtain graph partitioning. We define the distance measure as the inverse of the edge clustering coefficient. This coefficient is large when two nodes connected have many common neighbors.
For detecting communities, larger clustering coefficient will indicate smaller distance. In graphs, each node is characterized by the link density of its neighborhood. For generator points, we choose nodes with the highest local density within a region of a given radius.
Varying this radius allows community detection at different scales. Overall, the algorithm has three main steps. Assigning distances to edges, identifying generator nodes and constructing the Voronoi diagram. When applying the algorithm, the best strategy is to gradually increase the radius. For example, this is how the clustering changes in the American political Bloch sphere network as the radius grows.
We tested the algorithm on many other real and benchmark networks, also discussing its computational efficiency. In summary, our method shows that a topology-based metric on graphs can open new ways to approach graph community detection,
narrowing the gap between the relatively young field of community detection and the well-explored realm of cluster analysis in data mining.