Header menu link for other important links
X
Measuring the balance of signed networks and its application to sign prediction
, Bibhas Adhikari
Published in IOP PUBLISHING LTD
2017
Abstract
We propose a parametrized walk-based measure for the lack of balance in signed networks inspired by the Katz measure of similarity of two vertices in a network. We show that the performance of the proposed measure is marginally better than a recently proposed walk-based measure of the lack of balance for an undirected version of the real-world signed networks: Epinions, Slashdot and WikiElection. The proposed measure can be used to distinguish signed social networks on the basis of their degree of lack of balance. We also establish that cycles of shorter lengths can predict the sign of an edge in these signed networks better than the longer cycles by using the Katz prediction rule.
About the journal
JournalData powered by TypesetJOURNAL OF STATISTICAL MECHANICS-THEORY AND EXPERIMENT
PublisherData powered by TypesetIOP PUBLISHING LTD
ISSN1742-5468