Download presentation
Presentation is loading. Please wait.
1
Ranking in social networks
Miha Grčar Course in Knowledge Management Lecturer: prof. dr. Nada Lavrac Ljubljana, January 2007
2
Outline of the presentation
I. Prestige Structural prestige, social prestige, correlation Ways of calculating structural prestige II. Ranking Triad census Acyclic decomposition Symmetric-acyclic decomposition Ljubljana, January 2007 Miha Grčar
3
I. Prestige Prestigious people Structural prestige measures
People who receive many positive in-links Structural prestige measures Popularity or in-degree (Restricted) input domain Proximity prestige Structural prestige≠ social prestige (social status) Correlation between structural and social prestige Pearson’s correlation coefficient Spearman’s rank correlation coefficient Ljubljana, January 2007 Miha Grčar
4
Popularity or in-degree
1 2 3 1 Ljubljana, January 2007 Miha Grčar
5
Input domain Input domain size
How many nodes are path-connected to a particular node? Overall structure of the network is taken into account Problematic in a well-connected network 3 2 7 1 Ljubljana, January 2007 Miha Grčar
6
Restricted input domain
Resolves the input-domain issue in a well-connected network Issue: the choice of the maximum distance is quite arbitrary 3 2 5 1 Maximum distance = 2 Ljubljana, January 2007 Miha Grčar
7
Proximity prestige Eliminates the maximum-distance parameter
Closer neighbors are weighted higher 0.23 0.25 0.47 0.13 Input domain size / Number of nodes 7 / 8 Proximity prestige = = 0.47 Average path-distance to the node ( ( ) / 7 ( (3+3 Ljubljana, January 2007 Miha Grčar
8
II. Ranking We discuss techniques to extract discrete ranks from social relations Triad analysis helps us determine if our network is biased towards… Unrelated clusters (cluster = clique) Ranked clusters Hierarchical clusters Recipes for determining the hierarchy Acyclic decomposition Symmetric-acyclic decomposition Ljubljana, January 2007 Miha Grčar
9
Triad analysis Triads Atomic network structures (local)
16 different types M-A-N naming convention Mutual positive Asymmetric Null Ljubljana, January 2007 Miha Grčar
10
All 16 types of triads Ljubljana, January 2007 Miha Grčar
11
Triad census 6 balance-theoretic models Balance Clusterability
Ranked clusters Transitivity Hierarchical clusters (Theoretic model) Triad census: triads found in the network, arranged by the balance-theoretic model to which they belong Restricted to Unrelated clusters Ranked clusters Less and less restricted Hierarchical clusters Unrestricted Ljubljana, January 2007 Miha Grčar
12
Acyclic decomposition
Cyclic networks (strong components) are clusters of equals Acyclic networks perfectly reflect hierarchy Recipe Partition the network into strong components (i.e. clusters of equals) Create a new network in which each node represents one cluster Compute the maximum depth of each node to determine the hierarchy Ljubljana, January 2007 Miha Grčar
13
Acyclic decomposition An example
Ljubljana, January 2007 Miha Grčar
14
Acyclic decomposition An example
(1) (0) (2) The maximum depth of a node determines its position in the hierarchy Ljubljana, January 2007 Miha Grčar
15
Symmetric-acyclic decomposition
Strong components are not strict enough to detect clusters in the triad-census sense Symmetric-acyclic decomposition extracts clusters of vertices that are connected both ways After the clusters are identified, we can follow the same steps as in acyclic decomposition to determine the hierarchy Ljubljana, January 2007 Miha Grčar
16
Symmetric-acyclic decomp. An example
Ljubljana, January 2007 Miha Grčar
17
Reference Batagelj V., Mrvar A., de Nooy W. (2004): Exploratory Network Analysis with Pajek. Cambridge University Press Some figures used in the presentation are taken from this book Ljubljana, January 2007 Miha Grčar
Similar presentations
