Ego-zones: non-symmetric dependencies reveal network groups with large and dense overlaps

Milos Kudelka, Eliska Ochodkova, Sarka Zehnalova, Jakub Plesnik

SpringerOpen : Applied Network Science : Community structure in networks

VSB - Technical University of Ostrava, CS Dept.
17. listopadu 2172/15, 708 00
Czech Republic

Supplement

Example: Zachary's karate club

This example shows automatically detected prominent nodes and ego-zones in the network.
The algorithm published in the article above is based on the analysis of non-symmetric relationships between nodes.

Fig. 1: Three types on network nodes

Strongly (red), weakly (yellow), and non (green) prominent nodes are detected in the network.

Prominency in karate club

Fig. 2: Ego-zones generated by prominent egos

Overlapping and nested zones are detected around strongly, and weakly prominent nodes (red egos except node 31).
Zone generated by node 1 Zone generated by node 34
Zone generated by nodes 6 or 7 Zone generated by node 33
Zone generated by node 2 Zone generated by nodes 24 or 30
Zone generated by node 4 Zone generated by nodes 9 or 31
Zone generated by node 3 Zone generated by node 32

Fig 3: Ego-zones generated by non-prominent egos

A selection of additional zones; a total of 29 zones are detected in Zachary's karate club.

Zone generated by nodes 5 or 11 Zone generated by node 8
Zone generated by node 17 Zone generated by node 14
Zone generated by nodes 25 or 26 Zone generated by node 10

Milos Kudelka, 2020, milos.kudelka@vsb.cz