Network Biology (Sep 2014)

Using network properties to evaluate targeted immunization algorithms

  • Bita Shams,
  • Mohammad Khansari

Journal volume & issue
Vol. 4, no. 3
pp. 74 – 94


Read online

Immunization of complex network with minimal or limited budget is a challenging issue for research community. In spite of much literature in network immunization, no comprehensive research has been conducted for evaluation and comparison of immunization algorithms. In this paper, we propose an evaluation framework for immunization algorithms regarding available amount of vaccination resources, goal of immunization program, and time complexity. The evaluation framework is designed based on network topological metrics which is extensible to all epidemic spreading model. Exploiting evaluation framework on well-known targeted immunization algorithms shows that in general, immunization based on PageRank centrality outperforms other targeting strategies in various types of networks, whereas, closeness and eigenvector centrality exhibit the worst case performance.