Header menu link for other important links
Identifying influential spreaders in complex networks by weighted vote ranking and hybrid methods
S. Raamakirtinan,
Published in Little Lion Scientific
Volume: 99
Issue: 7
Pages: 1642 - 1661
Identifying influential users in complex networks has become the need of the hour in digital era and it has emerged as an important and interesting research topic. There are many methods like centrality measure, page rank, k-shell etc., to rank influential users, but each has its own shortfalls. In this paper we propose weighted vote ranking method and a hybrid ranking method as an extension of weighted vote ranking Method. The spreading ability of a node in a network mainly depend upon three key factors such as the support (vote) that it receives from neighbor nodes, the relationship strength (weight) and the nodes position in the network (coreness). So, to identify potential super spreaders we propose weighted vote ranking (WVR) method with all these three metrics such as vote, weight and coreness with a dynamic weight parameter control. Also, to improvise further we extended this proposed method by combining with weighted mixed degree decomposition (WMDD) method which has ability to consider removed and remaining nodes in decomposition technique. This hybrid method is obtained by combining these WVR and WMDD methods in 70:30 ratio which gives better results, compared to other ratio combinations. To study the spreading dynamics, Susceptible Infected Recovered (SIR) epidemic model is used. The average Kendall Tau of proposed weighted vote ranking and hybrid methods across various networks taken for experiment is 0.86 and 0.83 respectively whereas the highest average Tau from other methods is 0.79 (1 is maximum Tau value). Also, when we compared the average infection scale with other methods that is taken for experiment, we get 5.1%-14.5% higher infection scale for weighted vote ranking method and 8.8%-18.2% higher infection scale for hybrid method. Similar various other experiments like varying number of seed nodes and infection probability also show that proposed methods perform better when compared to recent other methods. © 2021 Little Lion Scientific.
About the journal
JournalJournal of Theoretical and Applied Information Technology
PublisherLittle Lion Scientific