1

Click here to load reader

Poster presented at EGC 2011

Embed Size (px)

Citation preview

Page 1: Poster presented at EGC 2011

Point of View Based Clustering ofSocio-Semantic NetworksInfluencing the communities dectection process in socio–semantic networks using points of view

Authors

Juan David CRUZ GOMEZCécile BOTHORELFrançois POULET

Partenaires

1 2

345

6

7

89

10 11 12 13

141516

17 18 19 20

2122

2324

25

26

27

29

28

1 2

345

6

7

89

10 11 12 13

141516

17 18 19 20

2122

2324

25

26

27

29

28

SemanticClustering

1 2

1 2

345

6

7

89

10 11 12 13

141516

17 18 19 20

2122

2324

25

26

27

29

28

1 2

345

6

7

89

10 11 12 13

141516

17 18 19 20

2122

2324

25

26

27

29

28

1

1

1

1

1

1

1

1

1

1

1

1

11

1

1

1

11

1

11

1

1

1

111

20

20

2020

20 2020

20

20

20

2020

20

3

Node 1

Node 2

Node 3

Node 4

Node 29

.

.

.

Feature 1 Feature 2 Feature 3

1

1

1

1 1

1

1 1

0 0

00

0

0

0

.

.

.

.

.

.

.

.

.

Point of View

1 2

345

6

7

89

10 11 12 13

141516

17 18 19 20

2122

2324

25

26

27

29

28

Social Graph

The point of viewis a set of binary vectors representinga subset of featuresfrom the socio-semantic network and assigned to each actorin it.

The social graph is therepresentation of therelationships betweenthe actors in the socio-semantic network.

Model Inputs Phase 1: semantic clustering

Phase 2: communities detection

The weights are changed according the semantic distance

1 2

345

6

7

89

10 11 12 13

141516

17 18 19 20

2122

2324

25

26

27

29

28

1 2

345

6

7

89

10 11 12 13

141516

17 18 19 20

2122

2324

25

26

27

29

28

4

The final communitites are structurally and semantically similar

Using Self-Organizing Maps [1] the nodes are clustered from a semantic perspective Each node belongs to one semantic group

3

4

1

2

The nodes semantically clustered according to the point of viewEach node in the network is assigned to one semantic group

The weights of the edges of the network are changed according to the semantic groups

The communities are found using the Fast Unfolding algorithm [2] on the social graph augmented withsemantic weights

Algorithm General Steps

References

[1] T. Kohonen, Self-Organizing Maps. Springer,1997.[2] V. D. Blondel, J.-L. Guillaume, R. Lambiotte, andE. Lefebvre, “Fast unfolding of communities in largenetworks,” Journal of Statistical Mechanics: Theoryand Experiment, vol. 2008, no. 10, p. P10008, 2008

Socio–semantic networks: enhancing the structurewith semantics

Inf ormat ion about the actorNameTypeDate of inclusion into the network

Social Graph InformationNode degree, node centrality, node betwenness, prestigeWalks and paths, relationships strenght, types of relationshipDensity of the graph, geodesics, distance and diameter, connectivity of the graph

This is the structural information the network

Semantic InformationRole of the actors, actor's name/filliation, actor's positionType of relationship, relationship statistics (date, evolution)Evolution of the network, contexts of the network updates, other features of the network

Points of view are created from these features

Socio-Semantic Network

By using the structural information and the semantical information in a conjointway it is possible to extract non–evident information and use it to analyze thenetwork from different perspectives.

Points of ViewGiven a graph G (V,E), let FV bethe set of semantic features of theactors of the network, and letF ∗V ∈ P (FV ) \FV , be a non–emptyset of features to be used todefine the point of view PoV .A point of view is defined as theset of all instances derived fromthe set FV :

PoVF ∗V =|V |⋃i=1ξi

where ξi is the binary vector(instance) assigned to the node i.

ConclusionAssigning weights derived from the results of the semantic clustering to the edges, the semanticinformation is included into the community detection process and the two types of data aremerged to find and visualize a social network from a selected point of view.

Contact : [email protected], http://www.telecom-bretagne.eu/