In this information age, the dissemination of information has a pivotal function in creating the adoption of new technologies, marketing of products, etc. However, the localization of information in a particular region is also a problem of the same importance to be addressed. For example, once a rumor is detected, it has to be localized at some parts of the network to obstruct its spreadability. This paper focuses on achieving information localization. The study says that localization can be achieved by manipulating the IPR value achievable through perturbation of the connections. In this paper, a Particle Swarm Optimization (PSO) based perturbation approach is proposed that passes through the steps of evolution to generate a graph with optimal IPR from the original graph. Hence, it guarantees a higher state of information localization in the output graph. The empirical results confirm a better localization by the proposed method than some existing methods. We use average optimal IPR as a parameter of evaluation.

错误:搜索内容不能为空,请输入英文关键词
错误:关键词超出字数限制,请精简
高级检索

Particle Swarm Optimization Based Graph Perturbation to Achieve Localization of Information Spread

  • Baishnobi Dash,
  • Debasis Mohapatra

摘要

In this information age, the dissemination of information has a pivotal function in creating the adoption of new technologies, marketing of products, etc. However, the localization of information in a particular region is also a problem of the same importance to be addressed. For example, once a rumor is detected, it has to be localized at some parts of the network to obstruct its spreadability. This paper focuses on achieving information localization. The study says that localization can be achieved by manipulating the IPR value achievable through perturbation of the connections. In this paper, a Particle Swarm Optimization (PSO) based perturbation approach is proposed that passes through the steps of evolution to generate a graph with optimal IPR from the original graph. Hence, it guarantees a higher state of information localization in the output graph. The empirical results confirm a better localization by the proposed method than some existing methods. We use average optimal IPR as a parameter of evaluation.